normal presentation of fetus

Appointments at Mayo Clinic

• Pregnancy week by week
• Fetal presentation before birth

The way a baby is positioned in the uterus just before birth can have a big effect on labor and delivery. This positioning is called fetal presentation.

Babies twist, stretch and tumble quite a bit during pregnancy. Before labor starts, however, they usually come to rest in a way that allows them to be delivered through the birth canal headfirst. This position is called cephalic presentation. But there are other ways a baby may settle just before labor begins.

Following are some of the possible ways a baby may be positioned at the end of pregnancy.

Head down, face down

When a baby is head down, face down, the medical term for it is the cephalic occiput anterior position. This the most common position for a baby to be born in. With the face down and turned slightly to the side, the smallest part of the baby's head leads the way through the birth canal. It is the easiest way for a baby to be born.

Head down, face up

When a baby is head down, face up, the medical term for it is the cephalic occiput posterior position. In this position, it might be harder for a baby's head to go under the pubic bone during delivery. That can make labor take longer.

Most babies who begin labor in this position eventually turn to be face down. If that doesn't happen, and the second stage of labor is taking a long time, a member of the health care team may reach through the vagina to help the baby turn. This is called manual rotation.

In some cases, a baby can be born in the head-down, face-up position. Use of forceps or a vacuum device to help with delivery is more common when a baby is in this position than in the head-down, face-down position. In some cases, a C-section delivery may be needed.

Frank breech

When a baby's feet or buttocks are in place to come out first during birth, it's called a breech presentation. This happens in about 3% to 4% of babies close to the time of birth. The baby shown below is in a frank breech presentation. That's when the knees aren't bent, and the feet are close to the baby's head. This is the most common type of breech presentation.

If you are more than 36 weeks into your pregnancy and your baby is in a frank breech presentation, your health care professional may try to move the baby into a head-down position. This is done using a procedure called external cephalic version. It involves one or two members of the health care team putting pressure on your belly with their hands to get the baby to roll into a head-down position.

If the procedure isn't successful, or if the baby moves back into a breech position, talk with a member of your health care team about the choices you have for delivery. Most babies in a frank breech position are born by planned C-section.

Complete and incomplete breech

A complete breech presentation, as shown below, is when the baby has both knees bent and both legs pulled close to the body. In an incomplete breech, one or both of the legs are not pulled close to the body, and one or both of the feet or knees are below the baby's buttocks. If a baby is in either of these positions, you might feel kicking in the lower part of your belly.

If you are more than 36 weeks into your pregnancy and your baby is in a complete or incomplete breech presentation, your health care professional may try to move the baby into a head-down position. This is done using a procedure called external cephalic version. It involves one or two members of the health care team putting pressure on your belly with their hands to get the baby to roll into a head-down position.

If the procedure isn't successful, or if the baby moves back into a breech position, talk with a member of your health care team about the choices you have for delivery. Many babies in a complete or incomplete breech position are born by planned C-section.

When a baby is sideways — lying horizontal across the uterus, rather than vertical — it's called a transverse lie. In this position, the baby's back might be:

• Down, with the back facing the birth canal.
• Sideways, with one shoulder pointing toward the birth canal.
• Up, with the hands and feet facing the birth canal.

Although many babies are sideways early in pregnancy, few stay this way when labor begins.

If your baby is in a transverse lie during week 37 of your pregnancy, your health care professional may try to move the baby into a head-down position. This is done using a procedure called external cephalic version. External cephalic version involves one or two members of your health care team putting pressure on your belly with their hands to get the baby to roll into a head-down position.

If the procedure isn't successful, or if the baby moves back into a transverse lie, talk with a member of your health care team about the choices you have for delivery. Many babies who are in a transverse lie are born by C-section.

If you're pregnant with twins and only the twin that's lower in the uterus is head down, as shown below, your health care provider may first deliver that baby vaginally.

Then, in some cases, your health care team may suggest delivering the second twin in the breech position. Or they may try to move the second twin into a head-down position. This is done using a procedure called external cephalic version. External cephalic version involves one or two members of the health care team putting pressure on your belly with their hands to get the baby to roll into a head-down position.

Your health care team may suggest delivery by C-section for the second twin if:

• An attempt to deliver the baby in the breech position is not successful.
• You do not want to try to have the baby delivered vaginally in the breech position.
• An attempt to move the baby into a head-down position is not successful.
• You do not want to try to move the baby to a head-down position.

In some cases, your health care team may advise that you have both twins delivered by C-section. That might happen if the lower twin is not head down, the second twin has low or high birth weight as compared to the first twin, or if preterm labor starts.

• Landon MB, et al., eds. Normal labor and delivery. In: Gabbe's Obstetrics: Normal and Problem Pregnancies. 8th ed. Elsevier; 2021. https://www.clinicalkey.com. Accessed May 19, 2023.
• Holcroft Argani C, et al. Occiput posterior position. https://www.updtodate.com/contents/search. Accessed May 19, 2023.
• Frequently asked questions: If your baby is breech. American College of Obstetricians and Gynecologists https://www.acog.org/womens-health/faqs/if-your-baby-is-breech. Accessed May 22, 2023.
• Hofmeyr GJ. Overview of breech presentation. https://www.updtodate.com/contents/search. Accessed May 22, 2023.
• Strauss RA, et al. Transverse fetal lie. https://www.updtodate.com/contents/search. Accessed May 22, 2023.
• Chasen ST, et al. Twin pregnancy: Labor and delivery. https://www.updtodate.com/contents/search. Accessed May 22, 2023.
• Cohen R, et al. Is vaginal delivery of a breech second twin safe? A comparison between delivery of vertex and non-vertex second twins. The Journal of Maternal-Fetal & Neonatal Medicine. 2021; doi:10.1080/14767058.2021.2005569.
• Marnach ML (expert opinion). Mayo Clinic. May 31, 2023.

Products and Services

• A Book: Mayo Clinic Guide to a Healthy Pregnancy
• 3rd trimester pregnancy
• Fetal development: The 3rd trimester
• Overdue pregnancy
• Pregnancy due date calculator
• Prenatal care: Third trimester

Mayo Clinic does not endorse companies or products. Advertising revenue supports our not-for-profit mission.

• Opportunities

Mayo Clinic Press

Check out these best-sellers and special offers on books and newsletters from Mayo Clinic Press .

• Mayo Clinic on Incontinence - Mayo Clinic Press Mayo Clinic on Incontinence
• The Essential Diabetes Book - Mayo Clinic Press The Essential Diabetes Book
• Mayo Clinic on Hearing and Balance - Mayo Clinic Press Mayo Clinic on Hearing and Balance
• FREE Mayo Clinic Diet Assessment - Mayo Clinic Press FREE Mayo Clinic Diet Assessment
• Mayo Clinic Health Letter - FREE book - Mayo Clinic Press Mayo Clinic Health Letter - FREE book
• Healthy Lifestyle

5X Challenge

Thanks to generous benefactors, your gift today can have 5X the impact to advance AI innovation at Mayo Clinic.

Warning: The NCBI web site requires JavaScript to function. more...

An official website of the United States government

The .gov means it's official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you're on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings
• Browse Titles

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing; 2024 Jan-.

StatPearls [Internet].

Delivery, face and brow presentation.

Julija Makajeva ; Mohsina Ashraf .

Affiliations

Last Update: January 9, 2023 .

• Continuing Education Activity

Face and brow presentation is a malpresentation during labor when the presenting part is either the face or, in the case of brow presentation, it is the area between the orbital ridge and the anterior fontanelle. This activity reviews the evaluation and management of these two presentations and explains the interprofessional team's role in safely managing delivery for both the mother and the baby.

• Identify the mechanism of labor in the face and brow presentation.
• Differentiate potential maternal and fetal complications during the face and brow presentations.
• Evaluate different management approaches for the face and brow presentation.
• Introduction

The term presentation describes the leading part of the fetus or the anatomical structure closest to the maternal pelvic inlet during labor. The presentation can roughly be divided into the following classifications: cephalic, breech, shoulder, and compound. Cephalic presentation is the most common and can be further subclassified as vertex, sinciput, brow, face, and chin. The most common presentation in term labor is the vertex, where the fetal neck is flexed to the chin, minimizing the head circumference. Face presentation is an abnormal form of cephalic presentation where the presenting part is the mentum. This typically occurs because of hyperextension of the neck and the occiput touching the fetal back. Incidence of face presentation is rare, accounting for approximately 1 in 600 of all presentations. [1] [2] [3]  In brow presentation, the neck is not extended as much as in face presentation, and the leading part is the area between the anterior fontanelle and the orbital ridges. Brow presentation is considered the rarest of all malpresentation, with a prevalence of 1 in 500 to 1 in 4000 deliveries. [3]

Both face and brow presentations occur due to extension of the fetal neck instead of flexion; therefore, conditions that would lead to hyperextension or prevent flexion of the fetal neck can all contribute to face or brow presentation. These risk factors may be related to either the mother or the fetus. Maternal risk factors are preterm delivery, contracted maternal pelvis, platypelloid pelvis, multiparity, previous cesarean section, and black race. Fetal risk factors include anencephaly, multiple loops of cord around the neck, masses of the neck, macrosomia, and polyhydramnios. [2] [4] [5]  These malpresentations are usually diagnosed during the second stage of labor when performing a digital examination. Palpating orbital ridges, nose, malar eminences, mentum, mouth, gums, and chin in face presentation is possible. Based on the position of the chin, face presentation can be further divided into mentum anterior, posterior, or transverse. In brow presentation, the anterior fontanelle and face can be palpated except for the mouth and the chin. Brow presentation can then be further described based on the position of the anterior fontanelle as frontal anterior, posterior, or transverse. Diagnosing the exact presentation can be challenging, and face presentation may be misdiagnosed as frank breech. To avoid any confusion, a bedside ultrasound scan can be performed. [6]  Ultrasound imaging can show a reduced angle between the occiput and the spine or the chin is separated from the chest. However, ultrasound does not provide much predictive value for the outcome of labor. [7]

• Anatomy and Physiology

Before discussing the mechanism of labor in the face or brow presentation, it is crucial to highlight some anatomical landmarks and their measurements. 

Planes and Diameters of the Pelvis

The 3 most important planes in the female pelvis are the pelvic inlet, mid-pelvis, and pelvic outlet. Four diameters can describe the pelvic inlet: anteroposterior, transverse, and 2 obliques. Furthermore, based on the landmarks on the pelvic inlet, there are 3 different anteroposterior diameters named conjugates: true conjugate, obstetrical conjugate, and diagonal conjugate. Only the latter can be measured directly during the obstetric examination. The shortest of these 3 diameters is obstetrical conjugate, which measures approximately 10.5 cm and is the distance between the sacral promontory and 1 cm below the upper border of the symphysis pubis. This measurement is clinically significant as the fetal head must pass through this diameter during the engagement phase. The transverse diameter measures about 13.5 cm and is the widest distance between the innominate line on both sides. The shortest distance in the mid pelvis is the interspinous diameter and usually is only about 10 cm. 

Fetal Skull Diameters

There are 6 distinguished longitudinal fetal skull diameters:

• Suboccipito-bregmatic: from the center of anterior fontanelle (bregma) to the occipital protuberance, measuring 9.5 cm. This is the diameter presented in the vertex presentation. 
• Suboccipito-frontal: from the anterior part of bregma to the occipital protuberance, measuring 10 cm 
• Occipito-frontal: from the root of the nose to the most prominent part of the occiput, measuring 11.5 cm
• Submento-bregmatic: from the center of the bregma to the angle of the mandible, measuring 9.5 cm. This is the diameter in the face presentation where the neck is hyperextended. 
• Submento-vertical: from the midpoint between fontanelles and the angle of the mandible, measuring 11.5 cm 
• Occipito-mental: from the midpoint between fontanelles and the tip of the chin, measuring 13.5 cm. It is the presenting diameter in brow presentation. 

Cardinal Movements of Normal Labor

• Neck flexion
• Internal rotation
• Extension (delivers head)
• External rotation (restitution)
• Expulsion (delivery of anterior and posterior shoulders)

Some key movements are impossible in the face or brow presentations. Based on the information provided above, it is obvious that labor be arrested in brow presentation unless it spontaneously changes to the face or vertex, as the occipito-mental diameter of the fetal head is significantly wider than the smallest diameter of the female pelvis. Face presentation can, however, be delivered vaginally, and further mechanisms of face delivery are explained in later sections.

• Indications

As mentioned previously, spontaneous vaginal delivery can be successful in face presentation. However, the main indication for vaginal delivery in such circumstances would be a maternal choice. It is crucial to have a thorough conversation with a mother, explaining the risks and benefits of vaginal delivery with face presentation and a cesarean section. Informed consent and creating a rapport with the mother is an essential aspect of safe and successful labor.

• Contraindications

Vaginal delivery of face presentation is contraindicated if the mentum is lying posteriorly or is in a transverse position. In such a scenario, the fetal brow is pressing against the maternal symphysis pubis, and the short fetal neck, which is already maximally extended, cannot span the surface of the maternal sacrum. In this position, the diameter of the head is larger than the maternal pelvis, and it cannot descend through the birth canal. Therefore, the cesarean section is recommended as the safest mode of delivery for mentum posterior face presentations. Attempts to manually convert face presentation to vertex, manual or forceps rotation of the persistent posterior chin to anterior are contraindicated as they can be dangerous. Persistent brow presentation itself is a contraindication for vaginal delivery unless the fetus is significantly small or the maternal pelvis is large.

Continuous electronic fetal heart rate monitoring is recommended for face and brow presentations, as heart rate abnormalities are common in these scenarios. One study found that only 14% of the cases with face presentation had no abnormal traces on the cardiotocograph. [8]  External transducer devices are advised to prevent damage to the eyes. When internal monitoring is inevitable, monitoring devices on bony parts should be placed carefully. 

Consultations that are typically requested for patients with delivery of face/brow presentation include the following:

• Experienced midwife, preferably looking after laboring women 1:1
• Senior obstetrician 
• Neonatal team - in case of need for resuscitation 
• Anesthetic team - to provide necessary pain control (eg, epidural)
• Theatre team  - in case of failure to progress, an emergency cesarean section is required.
• Preparation

No specific preparation is required for face or brow presentation. However, discussing the labor options with the mother and birthing partner and informing members of the neonatal, anesthetic, and theatre co-ordinating teams is essential.

• Technique or Treatment

Mechanism of Labor in Face Presentation

During contractions, the pressure exerted by the fundus of the uterus on the fetus and the pressure of the amniotic fluid initiate descent. During this descent, the fetal neck extends instead of flexing. The internal rotation determines the outcome of delivery. If the fetal chin rotates posteriorly, vaginal delivery would not be possible, and cesarean section is permitted. The approach towards mentum-posterior delivery should be individualized, as the cases are rare. Expectant management is acceptable in multiparous women with small fetuses, as a spontaneous mentum-anterior rotation can occur. However, there should be a low threshold for cesarean section in primigravida women or women with large fetuses.

The pubis is described as mentum-anterior when the fetal chin is rotated towards the maternal symphysis. In these cases, further descent through the vaginal canal continues, with approximately 73% of cases delivering spontaneously. [9]  The fetal mentum presses on the maternal symphysis pubis, and the head is delivered by flexion. The occiput is pointing towards the maternal back, and external rotation happens. Shoulders are delivered in the same manner as in vertex delivery.

Mechanism of Labor in Brow Presentation

As this presentation is considered unstable, it is usually converted into a face or an occiput presentation. Due to the cephalic diameter being wider than the maternal pelvis, the fetal head cannot engage; thus, brow delivery cannot occur. Unless the fetus is small or the pelvis is very wide, the prognosis for vaginal delivery is poor. With persistent brow presentation, a cesarean section is required for safe delivery.

• Complications

As the cesarean section is becoming a more accessible mode of delivery in malpresentations, the incidence of maternal and fetal morbidity and mortality during face presentation has dropped significantly. [10]  However, some complications are still associated with the nature of labor in face presentation. Due to the fetal head position, it is more challenging for the head to engage in the birth canal and descend, resulting in prolonged labor. Prolonged labor itself can provoke fetal distress and arrhythmias. If the labor arrests or signs of fetal distress appear on CTG, the recommended next step in management is an emergency cesarean section, which in itself carries a myriad of operative and post-operative complications. Finally, due to the nature of the fetal position and prolonged duration of labor in face presentation, neonates develop significant edema of the skull and face. Swelling of the fetal airway may also be present, resulting in respiratory distress after birth and possible intubation.

• Clinical Significance

During vertex presentation, the fetal head flexes, bringing the chin to the chest, forming the smallest possible fetal head diameter, measuring approximately 9.5 cm. With face and brow presentation, the neck hyperextends, resulting in greater cephalic diameters. As a result, the fetal head engages later, and labor progresses more slowly. Failure to progress in labor is also more common in both presentations compared to the vertex presentation. Furthermore, when the fetal chin is in a posterior position, this prevents further flexion of the fetal neck, as browns are pressing on the symphysis pubis. As a result, descending through the birth canal is impossible. Such presentation is considered undeliverable vaginally and requires an emergency cesarean section. Manual attempts to change face presentation to vertex or manual or forceps rotation to mentum anterior are considered dangerous and discouraged.

• Enhancing Healthcare Team Outcomes

A multidisciplinary team of healthcare experts supports the woman and her child during labor and the perinatal period. For a face or brow presentation to be appropriately diagnosed, an experienced midwife and obstetrician must be involved in the vaginal examination and labor monitoring. As fetal anomalies, such as anencephaly or goiter, can contribute to face presentation, sonographers experienced in antenatal scanning should also be involved in the care. It is advised to inform the anesthetic and neonatal teams in advance of the possible need for emergency cesarean section and resuscitation of the neonate. [11] [12]

• Review Questions
• Access free multiple choice questions on this topic.
• Comment on this article.

Disclosure: Julija Makajeva declares no relevant financial relationships with ineligible companies.

Disclosure: Mohsina Ashraf declares no relevant financial relationships with ineligible companies.

This book is distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0) ( http://creativecommons.org/licenses/by-nc-nd/4.0/ ), which permits others to distribute the work, provided that the article is not altered or used commercially. You are not required to obtain permission to distribute this article, provided that you credit the author and journal.

• Cite this Page Makajeva J, Ashraf M. Delivery, Face and Brow Presentation. [Updated 2023 Jan 9]. In: StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing; 2024 Jan-.

In this Page

Bulk download.

• Bulk download StatPearls data from FTP

Related information

• PubMed Links to PubMed

Similar articles in PubMed

• Sonographic diagnosis of fetal head deflexion and the risk of cesarean delivery. [Am J Obstet Gynecol MFM. 2020] Sonographic diagnosis of fetal head deflexion and the risk of cesarean delivery. Bellussi F, Livi A, Cataneo I, Salsi G, Lenzi J, Pilu G. Am J Obstet Gynecol MFM. 2020 Nov; 2(4):100217. Epub 2020 Aug 18.
• Review Sonographic evaluation of the fetal head position and attitude during labor. [Am J Obstet Gynecol. 2024] Review Sonographic evaluation of the fetal head position and attitude during labor. Ghi T, Dall'Asta A. Am J Obstet Gynecol. 2024 Mar; 230(3S):S890-S900. Epub 2023 May 19.
• Leopold Maneuvers. [StatPearls. 2024] Leopold Maneuvers. Superville SS, Siccardi MA. StatPearls. 2024 Jan
• Intrapartum sonographic assessment of the fetal head flexion in protracted active phase of labor and association with labor outcome: a multicenter, prospective study. [Am J Obstet Gynecol. 2021] Intrapartum sonographic assessment of the fetal head flexion in protracted active phase of labor and association with labor outcome: a multicenter, prospective study. Dall'Asta A, Rizzo G, Masturzo B, Di Pasquo E, Schera GBL, Morganelli G, Ramirez Zegarra R, Maqina P, Mappa I, Parpinel G, et al. Am J Obstet Gynecol. 2021 Aug; 225(2):171.e1-171.e12. Epub 2021 Mar 4.
• Review Labor with abnormal presentation and position. [Obstet Gynecol Clin North Am. ...] Review Labor with abnormal presentation and position. Stitely ML, Gherman RB. Obstet Gynecol Clin North Am. 2005 Jun; 32(2):165-79.

Recent Activity

• Delivery, Face and Brow Presentation - StatPearls Delivery, Face and Brow Presentation - StatPearls

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

Connect with NLM

National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894

Web Policies FOIA HHS Vulnerability Disclosure

Help Accessibility Careers

• Type 2 Diabetes
• Heart Disease
• Digestive Health
• Multiple Sclerosis
• Diet & Nutrition
• Health Insurance
• Public Health
• Patient Rights
• Caregivers & Loved Ones
• End of Life Concerns
• Health News
• Thyroid Test Analyzer
• Doctor Discussion Guides
• Hemoglobin A1c Test Analyzer
• Lipid Test Analyzer
• Complete Blood Count (CBC) Analyzer
• What to Buy
• Editorial Process
• Meet Our Medical Expert Board

Fetal Position in the Womb

• Risks and Complications
• Altering Fetal Position

Most fetuses are nestled inside the uterus (womb), curled up tight. This cozy position, knees to chest, is known as the fetal position. During pregnancy, the fetal position also refers to the direction a fetus faces in the uterus and is especially important as you approach delivery.

This article reviews the fetal position and how you and your providers change the fetal position before delivery when necessary.

Illustration by Zoe Hansen for Verywell Health

Fetal Position (or Presentation) In Utero

The ideal fetal position for birth is head down, spine parallel to the pregnant person's spine, face toward the back of the pregnant person's body with the chin tucked and arms folded across the chest. However, there are variations to the fetal position in utero that can affect delivery.

• Cephalic : The fetus is head down, with its chin tucked in and facing the pregnant person's spine.
• Breech : The fetus's buttocks or feet are toward the opening of the womb.
• Transverse : The fetus is sideways, at a 90-degree angle, to the pregnant person's spine.

Healthcare providers describe the fetal position in the uterus in terms of the fetal lie, position, and presentation.

Fetal lie refers to how the fetus's spine aligns with the gestational carrier's spine. Healthcare providers describe it as:

• Longitudinal : Parallel with the pregnant person's spine
• Transverse : Perpendicular to the pregnant person's spine
• Oblique : At an angle to the pregnant person's spine

Fetal Position

"Fetal position" refers to the direction the fetus is facing. The occipital bone is at the back of the fetus's head. Healthcare providers use this bone as a point of reference when describing fetal position, as follows: It is described as:

• Occiput anterior : The occipital bone is at the front of the birthing person's body, so the fetus is facing backward.
• Occiput posterior : The occipital bone is at the back of the birthing person's body, so the fetus is facing forward.

Fetal Presentation

Fetal presentation indicates the body part closest to the birth canal, also called the presenting part. The ideal presentation is the cephalic or vertex position. This when the fetus's head is down and the chin is tucked in and facing the spine. However, in some cases, the fetus can present with one of the following body parts closest to the birth canal:

• Buttocks (also known as the breech position)
• Face or brow

Positions and Risk of Delivery Complications 

Fetuses move, kick, and roll throughout pregnancy. However, during the third trimester, as space in the uterus gets tight, most fetuses naturally reposition into the cephalic fetal position, which is ideal for delivery.

However, some settle into breech or transverse positions. You can still deliver the baby in the following positions, but it can prolong labor and increase the risk of the following complications, which can restrict the baby’s oxygen supply:

• Shoulder dystocia : Occurs when the fetus's shoulder gets stuck in your pelvis
• Head entrapment : Occurs when the fetus's head is stuck inside a partially dilated cervix
• Umbilical cord compression or prolapse : Occurs when the umbilical cord is compressed and restricts oxygen and blood flow to the baby

How to Alter Fetal Position Before Delivery

When a healthcare provider performs an ultrasound and vaginal exam near the end of pregnancy , they may find that the fetus isn't in the ideal head-down position. They can help you explore options to alter the fetal position before delivery.

At home, you can try playing music by placing headphones or a speaker at the bottom of your uterus to encourage the fetus to turn. You can also put something cool on the top of your stomach and something warm (not hot) at the bottom to promote movement.

Specific exercises and yoga poses can help relax your pelvis and uterus, creating more room for the fetus and nudging it into the head-down position. Talk with your healthcare provider before attempting these techniques:  

• Cat-cow stretch : Get on your hands and knees and alternate between arching your back upward (like a cat) and dipping it downward (like a cow). 
• Pelvic circles : Gently make circles with your pelvis while standing.
• Child’s pose : Kneel on the ground, sit back on your heels, and stretch your arms forward, lowering your chest towards the ground. You can rest your forehead on the floor or on a cushion. Rest in this pose for 10-15 minutes. 
• Pelvic tilts : Lie on your back with your knees bent and your feet flat on the floor. Slowly tilt your hips upward, then release, returning to a neutral position. You can do this exercise for 10 to 20 minutes three times daily. 

Alternative options include seeing a chiropractor or acupuncturist that your healthcare provider recommends. Chiropractors align your hips and spine. Acupuncture is an Eastern medicine practice that involves inserting tiny needles in certain areas to balance your body’s energy. 

At the Hospital 

At the hospital, your provider may try an external cephalic version (ECV), in which they apply pressure to your belly to turn the fetus's head down.

Providers typically perform ECVs around 37 to 39 weeks' gestation, when the fetal size and the amount of amniotic fluid are ideal. An ECV is generally safe, but there are some risks, including fetal distress and preterm labor (rare).

The success rate of an ECV is about 60%. If an ECV is unsuccessful, your provider may recommend a surgical delivery known as a cesarean section (C-section). Before this surgical procedure, you will receive spinal anesthesia (numbing medicine), and your provider will make incisions in your belly to deliver the baby.

The fetal position indicates fetal alignment and presentation in the uterus. The cephalic (head-down) position is ideal for delivery. While it is possible to vaginally deliver a baby in other fetal positions, the risk of complications increases. There are ways to try to move the fetus at home or in the hospital; however, discuss these options with a healthcare provider before trying them at home.

Merck Manuals Consumer Version. Fetal presentation, position, and lie (including breech presentation) .

Yang L, Yi T, Zhou M, Wang C, Xu X, Li Y, Sun Q, Lin X, Li J, Meng Z. Clinical effectiveness of position management and manual rotation of the fetal position with a U-shaped birth stool for vaginal delivery of a fetus in a persistent occiput posterior position . J Int Med Res . 2020;48(6):300060520924275. doi:10.1177/0300060520924275

American Academy of Family Physicians. What can I do if my baby is breech ? 

Felemban AS, Arab K, Algarawi A, Abdulghaffar SK, Aljahdali KM, Alotaifi MA, Bafail SA, Bakhudayd TM. Assessment of the successful external cephalic version prognostic parameters effect on final mode of delivery . Cureus. 2021;13(7):e16637. doi:10.7759/cureus.16637

Angolile CM, Max BL, Mushemba J, Mashauri HL. Global increased cesarean section rates and public health implications: A call to action . Health Sci Rep . 2023;6(5):e1274. doi: 10.1002/hsr2.1274

By Brandi Jones, MSN-ED RN-BC Jones is a registered nurse and freelance health writer with more than two decades of healthcare experience.

Position and Presentation of the Fetus

• Fetal Presentation, Position, and Lie (Including Breech Presentation)

• About Merck
• Merck Careers

Toward the end of pregnancy, the fetus moves into position for birth. Normally, the position of a fetus is facing rearward (toward the woman’s back) with the face and body angled to one side and the neck flexed, and presentation is head first.

An official website of the United States government

Here’s how you know

Official websites use .gov A .gov website belongs to an official government organization in the United States.

Secure .gov websites use HTTPS A lock ( Lock Locked padlock icon ) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.

• Health Topics
• Drugs & Supplements
• Medical Tests
• Medical Encyclopedia
• About MedlinePlus
• Customer Support

Your baby in the birth canal

During labor and delivery, your baby must pass through your pelvic bones to reach the vaginal opening. The goal is to find the easiest way out. Certain body positions give the baby a smaller shape, which makes it easier for your baby to get through this tight passage.

The best position for the baby to pass through the pelvis is with the head down and the body facing toward the mother's back. This position is called occiput anterior.

Information

Certain terms are used to describe your baby's position and movement through the birth canal.

FETAL STATION

Fetal station refers to where the presenting part is in your pelvis.

• The presenting part. The presenting part is the part of the baby that leads the way through the birth canal. Most often, it is the baby's head, but it can be a shoulder, the buttocks, or the feet.
• Ischial spines. These are bone points on the mother's pelvis. Normally the ischial spines are the narrowest part of the pelvis.
• 0 station. This is when the baby's head is even with the ischial spines. The baby is said to be "engaged" when the largest part of the head has entered the pelvis.
• If the presenting part lies above the ischial spines, the station is reported as a negative number from -1 to -5.

In first-time moms, the baby's head may engage by 36 weeks into the pregnancy. However, engagement may happen later in the pregnancy, or even during labor.

This refers to how the baby's spine lines up with the mother's spine. Your baby's spine is between their head and tailbone.

Your baby will most often settle into a position in the pelvis before labor begins.

• If your baby's spine runs in the same direction (parallel) as your spine, the baby is said to be in a longitudinal lie. Nearly all babies are in a longitudinal lie.
• If the baby is sideways (at a 90-degree angle to your spine), the baby is said to be in a transverse lie.

FETAL ATTITUDE

The fetal attitude describes the position of the parts of your baby's body.

The normal fetal attitude is commonly called the fetal position.

• The head is tucked down to the chest.
• The arms and legs are drawn in towards the center of the chest.

Abnormal fetal attitudes include a head that is tilted back, so the brow or the face presents first. Other body parts may be positioned behind the back. When this happens, the presenting part will be larger as it passes through the pelvis. This makes delivery more difficult.

DELIVERY PRESENTATION

Delivery presentation describes the way the baby is positioned to come down the birth canal for delivery.

The best position for your baby inside your uterus at the time of delivery is head down. This is called cephalic presentation.

• This position makes it easier and safer for your baby to pass through the birth canal. Cephalic presentation occurs in about 97% of deliveries.
• There are different types of cephalic presentation, which depend on the position of the baby's limbs and head (fetal attitude).

If your baby is in any position other than head down, your doctor may recommend a cesarean delivery.

Breech presentation is when the baby's bottom is down. Breech presentation occurs about 3% of the time. There are a few types of breech:

• A complete breech is when the buttocks present first and both the hips and knees are flexed.
• A frank breech is when the hips are flexed so the legs are straight and completely drawn up toward the chest.
• Other breech positions occur when either the feet or knees present first.

The shoulder, arm, or trunk may present first if the fetus is in a transverse lie. This type of presentation occurs less than 1% of the time. Transverse lie is more common when you deliver before your due date, or have twins or triplets.

CARDINAL MOVEMENTS OF LABOR

As your baby passes through the birth canal, the baby's head will change positions. These changes are needed for your baby to fit and move through your pelvis. These movements of your baby's head are called cardinal movements of labor.

• This is when the widest part of your baby's head has entered the pelvis.
• Engagement tells your health care provider that your pelvis is large enough to allow the baby's head to move down (descend).
• This is when your baby's head moves down (descends) further through your pelvis.
• Most often, descent occurs during labor, either as the cervix dilates or after you begin pushing.
• During descent, the baby's head is flexed down so that the chin touches the chest.
• With the chin tucked, it is easier for the baby's head to pass through the pelvis.

Internal Rotation

• As your baby's head descends further, the head will most often rotate so the back of the head is just below your pubic bone. This helps the head fit the shape of your pelvis.
• Usually, the baby will be face down toward your spine.
• Sometimes, the baby will rotate so it faces up toward the pubic bone.
• As your baby's head rotates, extends, or flexes during labor, the body will stay in position with one shoulder down toward your spine and one shoulder up toward your belly.
• As your baby reaches the opening of the vagina, usually the back of the head is in contact with your pubic bone.
• At this point, the birth canal curves upward, and the baby's head must extend back. It rotates under and around the pubic bone.

External Rotation

• As the baby's head is delivered, it will rotate a quarter turn to be in line with the body.
• After the head is delivered, the top shoulder is delivered under the pubic bone.
• After the shoulder, the rest of the body is usually delivered without a problem.

Alternative Names

Shoulder presentation; Malpresentations; Breech birth; Cephalic presentation; Fetal lie; Fetal attitude; Fetal descent; Fetal station; Cardinal movements; Labor-birth canal; Delivery-birth canal

Barth WH. Malpresentations and malposition. In: Landon MB, Galan HL, Jauniaux ERM, et al, eds. Gabbe's Obstetrics: Normal and Problem Pregnancies . 8th ed. Philadelphia, PA: Elsevier; 2021:chap 17.

Kilpatrick SJ, Garrison E, Fairbrother E. Normal labor and delivery. In: Landon MB, Galan HL, Jauniaux ERM, et al, eds. Gabbe's Obstetrics: Normal and Problem Pregnancies . 8th ed. Philadelphia, PA: Elsevier; 2021:chap 11.

Review Date 11/10/2022

Updated by: John D. Jacobson, MD, Department of Obstetrics and Gynecology, Loma Linda University School of Medicine, Loma Linda, CA. Also reviewed by David C. Dugdale, MD, Medical Director, Brenda Conaway, Editorial Director, and the A.D.A.M. Editorial team.

Related MedlinePlus Health Topics

• Childbirth Problems

pregnancy health center / pregnancy a-z list / what are the different fetal positions article

What Are the Different Fetal Positions?

• Medical Author: Karthik Kumar, MBBS
• Medical Reviewer: Pallavi Suyog Uttekar, MD

5 Types of Fetal Positions and Presentations

• Comments **COMMENTSTAGLIST**
• More **OTHERTAGLIST**

The relationship between your baby's backbone and your backbone when your baby is in-utero is called the fetal position. Your baby can be in a variety of fetal positions, some make birth easier than others.

• Longitudinal position: The fetus’ and mother’s backbones are parallel to each other in this position.
• Transverse position: In this posture, the fetus’ backbone is at a right angle to the mother's backbone.
• Oblique position: The inclination angle of the fetus backbone is more than 0 and less than 90 degrees of the mother's backbone in this position.

Most people, however, confuse fetal position with the fetal presentation.

• Fetal position refers to whether the fetus is facing backward (facing the woman's back when she lies down) or forward (facing the woman's abdomen when she lies down).
• Fetal presentation is the body part of the baby that leads the way out of the birth canal.

The fetal position and presentation of your baby may influence the difficulty of your delivery. The baby may drop down into the pelvis before the due date. Here are some of the different positions and presentations your baby can get into while you are preparing for childbirth .

During pregnancy and when preparing for childbirth , there are exercises moms can do when the baby is active to get it in the optimal fetal position, which is known as baby spinning. Starting at the 35th week of pregnancy, talk to your doctor about maternal positioning.

Occiput anterior (OA) or vertex presentation

This is the optimal fetal positioning for childbirth . The baby enters the pelvis with their head down and chin tucked to the chest, facing the mother's back. The head points to the birth canal in this position.

There are two more presentations in OA:

• The baby will remain in the OA position, but the face, rather than the head, will be pointing toward the birth canal.
• This occurs when the chin is not tucked against the chest and instead points outward.
• During a vaginal examination, the doctor can detect this position by feeling the baby's bony jaws and mouth.
• In brow presentation, the baby will be in the OA position with their forehead pointing toward the birth canal. The doctor can feel the anterior fontanelle and the orbits of the forehead during the vaginal examination.
• One arm lies along with the head, pointing toward the birth canal.
• The arms may slide back during the delivery process, but if they do not, extra care must be taken to safely remove the baby.

Occiput posterior (OP)

• The baby enters the pelvis with its head down but facing the mother's front or abdomen.
• In general, approximately 10 to 34 percent of babies remain in the OP position during the first stage of labor before shifting to the optimal (OA) position.
• However, some babies remain in this position, which can make labor difficult and necessitate an emergency Cesarean delivery.
• This fetal position can cause labor to be prolonged, resulting in instrumental interventions, severe perineal tears, or Cesarean delivery.

The cephalic presentation or head-first positions are referred to as OA and OP.

Occiput transverse (OT)

• In the womb, the baby is lying sideways, and if they do not turn to the optimal position in time for birth, a Cesarean delivery is required.
• During a vaginal examination, the doctor may feel the shoulder, arm, elbow, or hand protruding into the vagina.
• This baby position increases the risk of umbilical cord prolapse, which occurs when the umbilical cord protrudes before the baby.
• Cord prolapse can occur in about one percent of babies in the transverse position, which is a medical emergency that necessitates an immediate Cesarean delivery.
• In some cases, assisted delivery is performed by manually rotating the baby or using forceps or a vacuum to position the baby in the ideal position.

Umbilical cord presentation

• During this time, the umbilical cord is the first to emerge from the birth canal.
• The condition of the uterine membrane, however, distinguishes umbilical cord presentation from prolapse.
• A cord presentation occurs when the umbilical cord enters the birth canal before the water breaks, whereas a cord prolapse occurs after the water breaks, necessitating an emergency Cesarean delivery.

Breech position

The infant is positioned with its buttocks directed toward the birth canal, resulting in the following types of breech positions:

• The buttocks are pointing toward the birth canal, with the legs folded at the knees and the feet close to the buttocks.
• In a vaginal delivery, this position increases the risk of an umbilical cord loop. Furthermore, the cord may pass through the cervix before the head, injuring the baby.
• The buttocks are pointing toward the birth canal with the legs straight up and the feet reaching the head.
• This can result in an umbilical cord loop, which can injure the baby during vaginal birth.
• The baby's buttocks are pointing down, and one of their feet is pointing toward the birthing canal.
• This can result in an umbilical cord prolapse, which can cut off the fetus' blood supply and oxygen supply.

A clinical examination of the abdomen, a vaginal examination, or an ultrasound examination is used to determine the position and presentation of the fetus during pregnancy.

Top What Are the Different Fetal Positions Related Articles

Childbirth Delivery Methods and Types

Newborn Skin Care

Labor Symptoms (Early Signs)

How Long Does It Take to Recover From Delivery?

Labor and Delivery

How Long Does It Take to Go Into Labor After Being Induced?

Why Is Normal Delivery Not Possible After Cesarean Delivery?

How long does it take to recover from a forceps delivery, what are the 4 stages of labor.

What Is the Process of Normal Delivery?

Why do doctors shave you before delivery of your baby.

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Data Descriptor
• Open access
• Published: 24 August 2024

A benchmark for 2D foetal brain ultrasound analysis

• Mariano Cabezas   ORCID: orcid.org/0000-0002-4417-1704 1   na1 ,
• Yago Diez 2   na1 ,
• Clara Martinez-Diago 3   na1 &
• Anna Maroto 3   na1  

Scientific Data volume  11 , Article number:  923 ( 2024 ) Cite this article

186 Accesses

Metrics details

• Scientific data

Brain development involves a sequence of structural changes from early stages of the embryo until several months after birth. Currently, ultrasound is the established technique for screening due to its ability to acquire dynamic images in real-time without radiation and to its cost-efficiency. However, identifying abnormalities remains challenging due to the difficulty in interpreting foetal brain images. In this work we present a set of 104 2D foetal brain ultrasound images acquired during the 20th week of gestation that have been co-registered to a common space from a rough skull segmentation. The images are provided both on the original space and template space centred on the ellipses of all the subjects. Furthermore, the images have been annotated to highlight landmark points from structures of interest to analyse brain development. Both the final atlas template with probabilistic maps and the original images can be used to develop new segmentation techniques, test registration approaches for foetal brain ultrasound, extend our work to longitudinal datasets and to detect anomalies in new images.

Similar content being viewed by others

Normative spatiotemporal fetal brain maturation with satisfactory development at 2 years

Automated body organ segmentation, volumetry and population-averaged atlas for 3D motion-corrected T2-weighted fetal body MRI

An automatic multi-tissue human fetal brain segmentation benchmark using the Fetal Tissue Annotation Dataset

Background & summary.

Foetal brain development is a complex sequence of events ocurrying throughout gestation. From early stages of embryonic development, the brain undergoes structural changes until several months after birth 1 , 2 . Therefore, understanding normal brain development is essential to identify potential deviations which may lead to neurological disability. Specifically, various studies have described normal milestones within a specific chronology 3 , 4 , 5 . These changes can be observed by experts in prenatal diagnosis with skills in foetal neurology using ultrasound (US) and magnetic resonance imaging (MRI). However, identifying neurodevelopmental deviations is challenging due to the difficulty in image interpretation 5 .

Currently, US is the established technique for screening due to its ability to acquire dynamic images in real-time without radiation and cost-efficiently. The international guidelines recommend the acquisition of a routine first trimester US within the 11th - 14th weeks and a mid-trimester scan within the 19th - 22th weeks of gestation for anatomical evaluation of the foetus 6 . The standard foetal assessment includes the evaluation of planes acquired using 2D-US. Even though 3D-US is considered useful for prenatal diagnosis of some disorders (mainly involving the face, the skeleton, the cardiovascular system or the brain), the major obstacles for 3D-US implementation worldwide as the main routine acquisition type are related to foetal motion artefacts and acoustic shadowing during volume acquisition. Another widely used prenatal imaging technique is MRI. However, MRI is not used as a primary screening tool and is instead used as a complementary acquisition when foetal abnormalities are suspected (even in selected high-risk cases) 7 , 8 . Moreover, foetal MRI performed before the 18th-22nd weeks does not usually provide additional information to that obtained by US. Generally, MRI provides a detailed visualization of structures between the 26th and 32nd weeks, being superior to US and less susceptible to limitations from maternal body conditions and foetal presentation (bones do not produce occlusion artefacts in MRI) 8 . Within this context, a foetal brain US atlas based on 3D scans of healthy foetuses was recently published 9 . This atlas demonstrates the feasibility of assessing structural changes in the cortex and in the subcortical grey matter by US. Furthermore, this chronological US-based atlas of the foetal brain at each gestational age has the potential to become a useful tool to detect development abnormalities when combined with the standard planes of mid-trimester routine scans. However, considering that the primary diagnostic tool in pregnancy remains 2D-US, studies aimed at improving the identification of fetal structures should focus on 2D-US.

As there is a degree of variability in the shapes of anatomical regions between individuals, atlases are typically built by taking into account a number of images considered to fall within normal parameters. A crucial step in this process is finding corresponding regions between different images and warping them to a common space, a process known as image registration. In the case of US atlases, the lower signal-to-noise ratio (SNR), the differences in location for the foetuses due to the lack of a robust localization technique and the presence of image artefacts that may blur or shadow salient features and structure boundaries makes this process particularly challenging.

In this paper we present a set of foetal brain 2D-US images with manual landmark annotations of structures of interest and soft probabilistic maps for those structures based on these landmarks in a common space. The aim of this dataset is to provide a set of US images as a starting point to study registration and segmentation of foetal brain scans and provide tools for researchers focusing on foetal brain development. Furthermore, in combination with the recently published 4D atlas, this dataset can be also used to study 2D to 3D US registration and techniques to determine the “real age” of gestation and potential abnormalities when comparing to the atlas. To define a common space, a registration algorithm based on fitting an ellipse to the skull segmentation of a convolutional neural network (CNN) was used to align all the images to a common space. In order to avoid misalignements caused by image quality, a subject ellipse was used to estimate an affine transformation to a common (atlas) space. We also took advantage of the automatic differentiation capabilities of the pytorch package in three fundamentally different optimization settings (training a CNN for segmentation, fitting an ellipse to a segmentation boundary, and estimating an affine transformation).

In order to create the dataset presented, the data collected was processed as follows: First, the skull was automatically segmented using a UNet network, then an ellipse was automatically fitted to the shape of the skull. These ellipses where used to register each image to a reference image. Soft probabilistic maps where built using the set of registered images. The rest of this section includes details on each of these steps and a summary of other publicly available datasets. A visual summary can be found in Fig.  1 .

Methodology: 1) The skull is automatically segmented using a UNet network (top-left). 2) An ellipse is fitted to the skull segmentation (top-right) 3) to estimate an affine transformation to a reference image (bottom-right). The axes of the fitted ellipse are warped to the image coordinate axes and are re-scaled to fit the ellipse in the reference image (bottom-left).

Ultrasound data

A prospective cohort of low-risk pregnant women was recruited at routine mid-trimester foetal ultrasound scan. All participants initiated antenatal care before the 12th weeks of gestation, underwent the first trimester ultrasound scan between 12th and 14th weeks and had a low risk for aneuploidies in the first trimester combined screening. Written informed consents were obtained from participants. A private dataset of 70 pregnant women with a routine mid-trimester foetal ultrasound scan at [20 ~ 20.6] weeks without detected abnormalities was acquired, totalling 104 scans (8 women were scanned three times, 18 women were scanned twice and the remaining 44 women were only scanned once). The median of the maternal age was 31 (range 18-42). Images were acquired using high-frequency transabdominal probe (C2-9) of Voluson E10 ultrasound system. For each subject, a transverse view of the foetal head demonstrating a standard normative transcerebellar scanning plane was manually annotated by2 trained clinical experts (Fig.  1 top right). Both experts were part of the Prenatal Diagnosis Unit, one with over 10 years of experience and the other with 5 years of experience. The experts jointly participated in the annotations of each imageto highlight common structures related to brain development as follows:

Skull : 4 landmarks from the inner line of the skull were annotated: 2 at the level of the middle line and the other 2 in a perpendicular imagined line at the level of the posterior corners of the cavum septi pellucidi (CSP).

Cerebellar peduncles (Thalami) : 1 landmark marking the edges of both thalami at the middle line and the outer edges of the concavity shape were annotated (3 total points).

Cerebellum : 8 landmarks on the perimeter the cerebellum were annotated. Specifically, 2 points from the midline, 2 points from the cerebellum external edges and 4 points in the middle of each cerebellum hemisphere.

Cavum septi pellucidi (Cavum) : 4 points, each marking one corner of its rectangular shape, were annotated.

Sylvian Fissure (Sylvius) : 2 landmarks, one for each sylvian fissure edge, and 1 landmark in the inflection point of the fissure were annotated. For all the images, only the inferior fissure was visible.

Midline : 1 landmark in the upper edge of the midline and 1 landmark in the upper edge of the CSP were annotated.

An important aspect of acquisition is that the sylvian fissure was always scanned on the lower part of the image, irrespective of the head orientation (left to right or right to left). This phenomena has important implications for registration. When aligning all images to a common 3D space (for example a 3D atlas template) in order to have all images facing the same direction, the transformation can be modeled with 180° rotations over the y axis. If the transformation is limited to the 2D space (image coordinates), a mirroring operation has to be applied to align all images. In the code we provide to process the images, the second option is used for simplicity. Furthermore, our common space is oriented from left to right (anterior to posterior).

Data acquisition was approved by the ethics committee “Comitè d’Ètica d’Investigació amb Medicaments CEIM GIRONA” with the code 2023.067. Furthermore, the subjects were informed and consented to the open publication of the data.

Skull segmentation database

The segmentation dataset used to train a skull segmentation CNN was downloaded from the HC18 grand challenge on “automated measurement of foetal head circumference using 2D ultrasound images” 10 , 11 . The challenge comprises a set of 800 × 540-pixel 2D US images with a pixel sizes ∈ (0.052, 0.326) mm. The data set was split into 999 images for training and 335 for testing. For each image in the training set, an ellipse was manually fitted to the HC by a trained sonographer but precise segmentations of the true skull boundary were not provided. No pre-processing techniques were applied to the images apart from computing the z-score of the intensity values with respect to the mean and standard deviation of the whole image (non-0 intensities) before feeding them to the segmentation network. The network trained with all the training set images was then used to roughly segment the skull of our own 2D US images and provide a starting point to fit an ellipse.

Other public databases

To the best of our knowledge, only 3 other large datasets focusing on 2-D ultrasound have been made publicly available, including HC18. As mentioned on our brief description of HC18, the annotations are limited (only an ellipse roughly representing the skull is given), the subjects present a large distribution of gestational ages and no further information on the acquired plan e is provided. For studies focusing on the segmentation of structures of interest or registration, new annotations would be needed. Furthermore, the gestational age and acquisition plane for each subject can have a large impact on the appearance of the structures. Another public dataset presented by Burgos-Artizzu et al . 12 includes a large set of images (12400) ranging from the 18th to the 40th week of pregnancy. Images were acquired using six different machines and labelled at the image level. Similarly to HC18, the dataset presents a large range of gestational ages and limited annotations that could only be used to develop classification algorithms. Finally, Alzubaidi et al . 13 released a public dataset of 3832 high resolution images. In contrast with the other two datasets, no mention of gestational age is provided and once again rough annotations in the form of bounding boxes are provided. Moreover, the authors also highlight image resampling as an additional shortcoming of their dataset.

In our dataset, we focus on a specific acquisition plane and gestational age as defined by international guidelines and raw images are provided. Furthermore, landmarks for the most salient points of each structure of interest are provided with software tools to estimate finer-grained masks and bounding boxes around the landmarks. In that sense, our dataset provides a useful tool to address multiple image analysis problems including registration that could not be easily approached with other available datasets and no additional annotations.

Registration method

Medical image registration 14 , 15 is a necessary initial step for many medical image processing applications that rely on group-wise analysis. Typically, images are registered in pairs: one of them is defined as the “fixed” image (or reference) and the other as the “moving” image. The moving image is then warped using a transformation function to generate the final “moved” image. The transformation is commonly optimised using a predefined metric that computes image similarity between the reference and the moved image after transformation. Here, we present a coarse registration method to roughly align different foetal ultrasound images as described in the following sections and illustrated in Fig.  1 . We chose one of the images as the common “reference” image (image 10) and registered the remaining images to it.

Automatic Skull segmentation using a Unet

The Unet architecture 16 is one of the most common CNN architectures for image segmentation. Due to its encoder-decoder structure and the use of skip connections, precise segmentations based on multi-scale features can be obtained for a variety of segmentation applications 17 . Consequently, we used a 2D Unet trained on the HC18 challenge dataset to provide a rough segmentation of the skull for all the images in our registration dataset (see Fig.  1 for an example). Specifically, the encoder and decoder blocks are comprised of 6 convolutional layers with a residual connection 18 of 32, 32, 128, 256, 256 and 1024 features each (inverse order for the decoder) and a bottleneck of 1024 features. To optimise the weights of the Unet, the Adam algorithm with default initial learning rate was used to minimise the binary cross-entropy loss.

Ellipse Registration

An ellipse is a planar curve representing the locus of the points with constant added distances to two “focal points”, as expressed by the quadratic equation:

With points that satisfy f E ( X ,  Y ) ≠ 0 being inside the ellipse perimeter ( f E ( X ,  Y ) < 0) or outside of it ( f E ( X ,  Y ) > 0). The general equation’s coefficients can be obtained from known semi-major axis a (represented by the magnitude of the turquoise vector in Fig.  1 ), semi-minor axis b (represented by the magnitude of the cyan vector in Fig.  1 ), centre coordinates ( x 0 ,  y 0 ) (represented by the point where the two vectors meet in Fig.  1 ) and rotation angle θ (the angle from the positive horizontal axis to the ellipse’s major axis as observed in Fig.  1 ) using the formulae:

These expressions can be derived from the canonical equation \(\frac{{x}^{2}}{{a}^{2}}+\frac{{y}^{2}}{{b}^{2}}=1\) by an affine transformation of the coordinates ( x ,  y ) (with a translation (− x 0 , − y 0 ) and an angle θ ).

Given the set of pixel coordinates of the skull segmentation ( \(({X}_{skull},{Y}_{skull})=\,[({x}_{1},{y}_{1}),\ldots ,({x}_{N},{y}_{N})]\) ) we can fit an ellipse using its parameters ( a , b , x 0 , y 0 and t h e t a ) by minimising the following objective function:

where the coefficients A to F are substituted in f E by their definitions in equations ( 2 ) to ( 7 ) and the estimated ellipse parameters.

This process is repeated 5 times, removing erroneous points of the skull segmentation mask that are one standard deviation away from the mean ellipse error according to Eq. ( 10 ). With the parameters of the skull ellipse estimated through optimisation, we can now define an affine transformation (referred to as Ellipse from now on) to move the brain to the centre of the image as:

Affine image registration

For comparison, a regular rigid registration of 6 unrestricted parameters was performed with different initialisations. From the most basic identity initialisation (referred to as Affine ), to an initialisation using the ellipse parameters of the reference image (referred to as Affine (Reference ellipse) ) and a refinement of the ellipse-based affine transformation from Eq. ( 11 ) (referred to as Ellipse + Affine ).

Probabilistic maps

Once the images are co-registered to a common space based on their ellipse, a two-step process is performed to generate probabilistic maps for all the structures with more than 2 landmarks.

First, the concave hull of each structure is computed using the alphashapes package. This concave hull represents a rough segmentation of the location of the structure. Second, the segmentations for all the subjects are averaged per structure to determine the probability of each pixel to belong to that structure.

This approach has some limitations. Namely, some structures have a polygonal shape, even though they are actually curves (e.g. sylvius) and the final masks are only a rough representation of the real boundaries (e.g. cerebellum). However, these polygonal maps can be still used to determine growth milestones and to provide a rough location of the structure of interest.

Data Records

The original images with manual landmark annotations (Gimp image editing tool format) and the co-registered images and probabilistic maps used within this paper can be found on figshare 19 , and are organised with subject id number (1-52) and an additional number for multiple scans (e.g. 36 and 36.1). Co-registered images are saved in JPEG format with the “_registered” suffix. The final probability maps estimated using a combination of co-registered landmarks and the alphashape package are provided as JPEG images with the name of the structure (e.g. sylvius.jpeg), while comma separated value (CSV) files with the point landmarks of all registered subjects are compressed into a single zip file.

In total, the released dataset consists of 104 annotated 2D US images of foetal brains on the 20th week of pregnancy. The manual annotations are described in Ultrasound dataset section.

Technical Validation

To validate the techniques used for co-registration to a common space, we focused on common medical imaging metrics for registration using landmarks. In this section we describe these metrics and provide some qualitative and quantitative results of the alignment (including a visualisation of the probabilistic maps of one of the structures of interest).

Regarding the quality of the US images, all images were acquired with a high frequency ultrasound probe (2-9 MHz) that provides high resolution images following the ISUOG recommendations. The most constraining factor for the quality of the ultrasound images is maternal obesity. However, for this study we excluded women with maternal morbid obesity (body mass index > 40) as it is one of the factors for high-risk pregnancy. To further illustrate that point, we provide a comparative example between an image from the dataset and a lower quality one in Fig.  2 .

Qualitative comparison between a low quality image where structures are not clearly visible and and image from the dataset.

Evaluation metrics

For this study we chose to use the anatomical knowledge defined by expert annotations as the main directive to evaluate the quality of the registration and avoid focusing exclusively on common pixel-metrics that might be unreliable and disconnected from physical properties 20 . Specifically, we used the point annotations described in the Ultrasound Data section with two point-based metrics and one area-based metric for every anatomical structure with more than 2 points (all the structures except the midline). For completeness sake, we also included the structural similarity index metric (SSIM) pixel-based metric.

Point-based metrics

In order to penalise partial matches between anatomical structures, we considered the Hausdorff distance ( d H ), that computes the worst possible Euclidean distance ( d ( p i ,  p j )) between two sets of points P f (| P f | =  N ) and P m (| P m | =  M ). We also considered the average of the minimum Euclidean distances ( d E ) to express global similarity between structures defined by landmarks.

Area-based metric

To evaluate the superposition between two anatomical structures defined as 2D landmarks (points), we first computed their concave hulls and then considered their Dice similarity coefficient (polygon DSC).

Image similarity metric

For completeness, we also considered the structural similarity index measure (SSIM) as a pixel-intensity-based metric.

Comparison of coarse registration approaches

Figures  3 , 4 and 5 summarise the results with boxplots and Wilcoxon signed-rank tests for all the registration methods and metrics considered. Wherever possible, the results for different anatomical areas are presented separately. For the Euclidean and Hausdorff metrics, lower values indicate better registration, while for the SSIM and Dice metric higher results indicate better registration. Moreover, a qualitative example to illustrate misalignments between the reference points and the registered landmarks is provided in Fig.  6 .

Quantitative results for all the methods on each structure (Hausdorff distance, d H , lower values indicate better registration). The upper part of each boxplot figure indicates the results of pairwise statistical Wilcoxon tests: (ns: 5.00e-02  < p ≤ 1.00e+00, * 1.00e-02  < p ≤ 5.00e-02, ** 1.00e-03  < p ≤ 1.00e-02, *** 1.00e-04  < p ≤ 1.00e-03, **** p ≤ 1.00e-04).

Quantitative results for all the methods on each structure with multiple points (polygon DSC, higher values indicate better registration). The upper part of each boxplot figure indicates the results of pairwise statistical Wilcoxon tests: (ns: 5.00e-02  < p ≤ 1.00e+00 * 1.00e-02  < p ≤ 5.00e-02, ** 1.00e-03  < p ≤ 1.00e-02, *** 1.00e-04  < p ≤ 1.00e-03, **** p ≤ 1.00e-04).

Quantitative results for all the methods according to the SSIM metric (higher values indicate better registration). The upper part of each boxplot figure indicates the results of pairwise statistical Wilcoxon tests: (ns: 5.00e-02  < p ≤ 1.00e+00, * 1.00e-02  < p ≤ 5.00e-02, ** 1.00e-03  < p ≤ 1.00e-02, *** 1.00e-04  < p ≤ 1.00e-03, **** p ≤ 1.00e-04).

Qualitative example of the alignment between the reference image and a randomly selected subject (11). The background image corresponds to the warped image after registration, circles denote the reference points for each structure, crosses represent the registered points and lines are used to illustrate misalignment between the reference and the registration.

Regarding point metrics, the ellipse ( E ) and ellipse with affine methods ( E+A ) obtain overall better results than the other methods. In general, the differences observed were found to be statistically significant for both point metrics and most anatomical structures. Exceptions to this are the thalami and the cerebellum where the pixel based affine registration method initialised using the reference ellipse ( AFF+I ), obtains results that appear worse but are not significantly different. Comparing the E and E+A methods, small (and not statistically significant) differences can be observed. Using a refinement registration after ellipse-based method slightly worsens the metrics when applied to the skull but improves them in all other anatomical structures. This is an expected result as the main focus of the first method is to co-register the skulls (ellipses). The uninitialised affine transformation ( AFF ) obtains significantly worse results than these two methods and has a higher variance of values as illustrated by its wide boxplots. Both methods using exclusively pixel-wise affine registration (AFF, AFF+I) achieve results that are worse than the metrics of the original moving image in some cases. This illustrates the disconnect between pixel-based metrics guiding these methods and point-based metrics targeting distances between real anatomical structures (especially for noisy sequences).

Regarding the differences between the Euclidean and Hausdorff metrics (especially the median values shown by the central line in boxplots), slightly higher Hausdorff values indicate that some point pairs are further than the average euclidean distance mean value. This is especially noticeable on the skull for E and A+E and the thalami for AFF .

The other two metrics show the same general tendencies, even though they focus on different aspects of the registration. The low SSIM values observed for all methods (even though E and A+E outperform all methods) illustrats how challenging this registration scenario is. The variations between individuals and acquisitions and low SNR make the intensity values particularly unreliable. On the other hand, the high anatomical DSC results observed for the ellipse-based methods (includding AFF+I) validate our geometric approach that relies on a rough initial skull estimate to determine the general shape and orientation of the brain. Similarly high results obtained for the cerebellum indicate that a correct skull placement is a crucial important step towards the registration of all brain structures.

Finally, Fig.  7 shows the final heatmap generated from the registered concave hulls (E+A) of the cerebellum for all output images. Even though faint outlines of some incorrect registration results outside of the cerebellum region (delimited by orange points) can be observed, the higher probability regions in the heatmap clearly correspond to the cerebellum region of the reference image delimited by the manually annotated orange landmarks.

Probabilistic map for the cerebellum structure based on the averaged concave hull of the landmarks for each image. Landmarks for the reference image (background) are also provided.

Usage Notes

The code provided to analyse the images and perform a rough initial alignment to a common space has been developed using python. A set of Jupyter notebooks detailing the use of each step is also provided in the repository with visualisation examples of each step of the processing pipeline. Furthermore, we provide the trained weights for the skull segmentation network as part of the data records (file unet.pt 19 ).

Regarding the use of python packages, the code heavily relies on pytorch (version 1.12.0 with CUDA 10.2) to do the heavy lifting. However, numpy (version 1.21.6), scipy (version 1.7.3), scikit-learn (version 1.0.2), scikit-image (version 0.18.1), pandas (version 1.3.4), persim (version 0.3.1) and alphashape (version 1.3.1) are also used for some of the processing steps or to analyse different metrics related to the registration. Furthermore, opencv-python (version 4.5.2.52) and gimpformats (version 2021.1.4) where used to open the images with python. In particular, manual landmark annotation where done using the Gimp software and were stored (together with the ultrasound image) using the xcf format. Finally, matplotlib (version 3.7.1), seaborn (version 0.11.0) and statannot (version 0.2.3) were used for results visualisation inside the Jupyter notebooks.

Code availability

All the code used in the study to generate the final atlas and to co-register the images is publicly available at https://github.com/marianocabezas/fetal-brain .

Monteagudo, A. & Timor-Tritsch, I. E. Normal sonographic development of the central nervous system from the second trimester onwards using 2d, 3d and transvaginal sonography. Prenatal Diagnosis 29 , 326–339 (2009).

Article   PubMed   Google Scholar  

Namburete, A. I. et al . Learning-based prediction of gestational age from ultrasound images of the fetal brain. Medical Image Analysis 21 , 72–86 (2015).

Article   PubMed   PubMed Central   Google Scholar  

Cohen-Sacher, B., Lerman-Sagie, T., Lev, D. & Malinger, G. Sonographic developmental milestones of the fetal cerebral cortex: a longitudinal study. Ultrasound in Obstetrics & Gynecology 27 , 494–502 (2006).

Article   CAS   Google Scholar  

Kline-Fath, B. M. Ultrasound and mr imaging of the normal fetal brain. Neuroimaging Clinics of North America 29 , 339–356 (2019).

Guimaraes, C. V. & Dahmoush, H. M. Fetal brain anatomy. Neuroimaging Clinics of North America 32 , 663–681 (2022).

Salomon, L. J. et al . Isuog practice guidelines (updated): performance of the routine mid-trimester fetal ultrasound scan. Ultrasound in Obstetrics & Gynecology 59 , 840–856 (2022).

Griffiths, P. D. et al . Mri in the diagnosis of fetal developmental brain abnormalities: the meridian diagnostic accuracy study. Health Technology Assessment (Winchester, England) 23 , 1 (2019).

Prayer, D. et al . ISUOG practice guidelines (updated): performance of fetal magnetic resonance imaging. Ultrasound in Obstetrics & Gynecology 61 , 278–287 (2023).

Namburete, A. I. et al . Normative spatiotemporal fetal brain maturation with satisfactory development at 2 years. Nature 1–9 (2023).

van den Heuvel, T. L. A., de Bruijn, D., de Korte, C. L. & van Ginneken, B. Automated measurement of fetal head circumference using 2d ultrasound images. PloS one 13 , e0200412 (2018).

van den Heuvel, T. L. A., de Bruijn, D., de Korte, C. L. & van Ginneken, B. Automated measurement of fetal head circumference using 2d ultrasound images [data set]. Zenodo , https://doi.org/10.5281/zenodo.1322001 (2024).

Burgos-Artizzu, X. P. et al . Evaluation of deep convolutional neural networks for automatic classification of common maternal fetal ultrasound planes. Scientific Reports 10 , 10200 (2020).

Article   ADS   CAS   PubMed   PubMed Central   Google Scholar  

Alzubaidi, M. et al . Large-scale annotation dataset for fetal head biometry in ultrasound images. Data in Brief 51 , 109708 (2023).

Article   CAS   PubMed   PubMed Central   Google Scholar  

Fu, Y. et al . Deep learning in medical image registration: a review. Physics in Medicine & Biology 65 , 20TR01 (2020).

Article   Google Scholar  

Gholipour, A., Kehtarnavaz, N., Briggs, R., Devous, M. & Gopinath, K. Brain functional localization: A survey of image registration techniques. IEEE Transactions on Medical Imaging 26 , 427–451, https://doi.org/10.1109/TMI.2007.892508 (2007).

Ronneberger, O., Fischer, P. & Brox, T. U-net: Convolutional networks for biomedical image segmentation. In Proceedings of the International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI) , 234–241 (2015).

Isensee, F., Jaeger, P. F., Kohl, S. A. A. & Maier-Hein, K. H. nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation. Nature: Methods (2021).

He, K., Zhang, X., Ren, S. & Sun, J. Deep residual learning for image recognition. In IEEE Conf. Comput. Vision Pattern Recog ., 770 – 778 (2016).

Cabezas, M., Diez, Y., Martinez-Diago, C. & Maroto, A. A benchmark for 2d foetal brain ultrasound analysis [data set]. Figshare , https://doi.org/10.6084/m9.figshare.c.6984822.v1 (2024).

Rohlfing, T. Image similarity and tissue overlaps as surrogates for image registration accuracy: Widely used but unreliable. IEEE Transactions on Medical Imaging 31 , 153–163 (2012).

Download references

Author information

These authors contributed equally: Mariano Cabezas, Yago Diez, Clara Martinez-Diago, Anna Maroto.

Authors and Affiliations

Brain and Mind Centre, University of Sydney, Sydney, Australia

Mariano Cabezas

Faculty Of Science, Yamagata University, Yamagata, Japan

Hospital Universitari de Girona Doctor Josep Trueta, Girona, Spain

Clara Martinez-Diago & Anna Maroto

You can also search for this author in PubMed   Google Scholar

Contributions

All authors conceived and designed the experiments, C.M. and A.M. collected the images and performed the landmark annotations, M.C. and Y.D. conducted the experiments and analysed the results. All authors reviewed the manuscript.

Corresponding author

Correspondence to Mariano Cabezas .

Ethics declarations

Competing interests.

The authors declare no competing interests.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/ .

Reprints and permissions

About this article

Cite this article.

Cabezas, M., Diez, Y., Martinez-Diago, C. et al. A benchmark for 2D foetal brain ultrasound analysis. Sci Data 11 , 923 (2024). https://doi.org/10.1038/s41597-024-03774-3

Download citation

Received : 27 December 2023

Accepted : 14 August 2024

Published : 24 August 2024

DOI : https://doi.org/10.1038/s41597-024-03774-3

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Sign up for the Nature Briefing: AI and Robotics newsletter — what matters in AI and robotics research, free to your inbox weekly.