2025 Volume 13 Issue 2 Pages 39-45
Aim: To evaluate the utility of a three-dimensional (3D) labor simulation technique based on low-dose computed tomography.
Methods: We obtained computed tomography data for an expectant mother who required pelvimetry to diagnose cephalopelvic disproportion at 36 weeks’ gestation. Using low-dose computed tomography data, 3D models of the maternal pelvis and the fetal head were created, based on which simulations to predict labor progression were performed. The patient provided informed consent for the use of her data for research purposes.
Results: Our low-dose computed tomography protocol has been demonstrated to result in lower radiation exposure (0.17 mGy) than conventional X-ray pelvimetry (0.50–1.18 mGy). Computed tomography images clearly visualized the pelvic and fetal head dimensions. The 3D cephalopelvic model effectively demonstrated the spatial relationship between the fetal head and the maternal pelvis, allowing for a comprehensive assessment that could simulate labor progression.
Conclusions: This novel 3D simulation system based on low-dose computed tomography not only offers a minimally invasive and informative means to predict labor progression but also provides a visual and tactile method for assessing fetal progression through the birth canal, potentially improving the accuracy of prediction of cephalopelvic disproportion. Further large-scale studies will be necessary to validate the clinical efficacy and safety of this 3D simulation system.
Obstructed or pathological labor due to cephalopelvic disproportion (CPD) is a serious health problem that can result in maternal and fetal injuries or even life-threatening complications. Morbidities associated with labor obstruction range from acute complications such as uterine rupture and chorioamnionitis to chronic complications such as uterine fistula formation, maternal incontinence, and neurological damage to the offspring. According to recent reports, the incidence of CPD is 1%–8%,1) and an estimated >1 million pregnant women suffer from labor obstruction worldwide.2) CPD is one of the leading indications for emergency cesarean section, as precisely predicting CPD before labor begins is difficult. In other words, if CPD could be predicted before labor onset, the risk of emergency cesarean section would be avoided, preventing unnecessary labor trials. To this end, there is a need to develop clinical methods for diagnosing CPD.
Classically, the true obstetric conjugate, i.e., anteroposterior diameter of the pelvic inlet (APDPI), is measured using X-ray pelvimetry.3) However, performing conventional radiographic pelvimetry is challenging for the following reasons. First, it is difficult to obtain clear images. Second, it exposes the fetus to relatively high amounts of radiation.4) Data from atomic bomb survivors suggest that the minimal threshold for adverse effects on the fetus for congenital anomalies, growth restriction, and abortion is in the range of 60–310 mGy. Since current radiation imaging involves much lower radiation exposure, the recent American College of Obstetricians and Gynecologists guidelines state that radiation exposure from radiography and computed tomography (CT) is much lower than that associated with fetal harm.5)
CT has been introduced as a substitute for conventional methods,6,7,8) and in particular, low-dose CT has been used as a validated method that results in significantly lower fetal radiation exposure, yet superior image quality, compared with conventional X-ray pelvimetry. Nevertheless, with a recent systematic review failing to demonstrate the ability of pelvimetry to predict labor obstruction,9) a more effective method of diagnosing CPD is needed.
Three-dimensional (3D) modeling technology was recently introduced in orthopedic surgery. For example, mapping technology using a 3D patient model is gaining popularity for the preoperative planning of fracture surgeries.10) In the present study, we created 3D models of a patient’s pelvis and the fetal head based on CT data, and developed a simulation system using these models that would provide a new technique for simulating labor progression.
The Nerima Hikarigaoka Hospital Institutional Review Board approved this study.
Patient background and clinical courseA 32-year-old woman (gravida 2, para 1) visited our hospital at 11 weeks and 2 days of gestation. She had previously delivered a baby weighing 2,644 g. Fearing labor obstruction, she requested pelvimetry before the next delivery. The potential benefits of an improved CPD diagnosis were carefully weighed against the minimal radiation risk and deemed to outweigh the potential harm. Upon receiving a thorough explanation of the procedure and potential fetal radiation risk, the patient underwent CT pelvimetry at a gestational age of 36 weeks and 1 day. She provided consent for the use of her CT data for research purposes.
Computed tomographyPrior to the introduction of CT pelvimetry in our hospital, we performed a preliminary estimation of fetal radiation exposure. Following the “as low as reasonably achievable” (ALARA) principle, we optimized CT parameters to minimize fetal radiation exposure while maintaining diagnostic image quality.11) Images were obtained with an Aquilion ONE NATURE Edition 320-row CT scanner (CANON Medical Systems Co.). A CT Colonography Phantom NCCS (Kyoto Kagaku Co.) was used to estimate fetal radiation exposure, with a dosimeter inserted into the phantom representing the maternal pelvis and fetus. The phantom was placed between two acrylic boards. The following CT settings were used: 120 kV, scanning speed 0.5 r/s, pitch factor 0.813, beam width 4.0 cm, and calibration 9.7. Deep learning reconstruction (AiCE; CANON Medical Systems Co.) was used to reduce image noise. The exposure dose was measured at a site corresponding to the fetal head and compared with doses of conventional X-ray pelvimetry using Model 0.6 (Shimadzu Medical Systems Co.). Estimated exposure doses were 0.24 mGy for the Martius method, 0.26 mGy for the Guthman method, and 0.17 mGy (10 mAs) and 0.34 mGy (20 mAs) for low-dose CT.
CT pelvimetryAt a gestational age of 36 weeks and 1 day, the patient underwent a low-dose CT examination of the pelvis while in the supine position. Her height was 149 cm, body weight was 45.75 kg, and body mass index was 20.6. The estimated fetal weight on ultrasound (Shinozuka method) was 2,316 g immediately before the examination. Images were acquired with the CT settings maintained at 120 kV and 10 mA. Collected data were analyzed by multiple radiologists. The obstetric diameters of the pelvis and fetal head were also measured. To assess method reproducibility, measurements were repeated for each diameter by different examiners.
An image of the pelvic inlet plane (Martius’ method) was obtained. The pelvic inlet shape was labeled according to the Caldwell and Moloy classification.12,13) A pelvic inlet plane image with a transverse section of the fetal head was analyzed.14) The diameters were measured and evaluated (Figures 1A and 1B) for clinical purposes.
The APDPI (corresponding to the true obstetric conjugate) was 11.4 cm, and the anteroposterior diameter of the pelvic outlet (APDPO) was 11.7 cm. The coccygeal pelvic outlet was 11.6 cm, the transverse inlet diameter (TID) was 12.3 cm, the bispinous diameter was 10.3 cm, and the biparietal diameter (BPD) of the fetus was 9.2 cm. All diameters showed good reproducibility and inter-observer agreement (data not shown). The diagnosis of contracted pelvis for Japanese woman involves an APDPI ≤9.5 cm or TID ≤10.5 cm; patients with an APDPI >9.5 cm, but TID ≤10.5 cm, are diagnosed with relatively contracted pelvis.15) The risk of obstructed labor has been reported to increase with APDPO ≤10.7 cm.16) Based on these criteria, the patient was identified as not having a contracted pelvis.
The cephalopelvic disproportion index (difference between APDPI and BPD) was 2.1 cm, suggesting the absence of CPD.17) Additionally, a two-dimensional evaluation was performed in which all five examiners independently diagnosed the patient’s pelvic inlet shape as gynecoid/gynecoid (anterior/posterior). All of the examiners determined that the fetal head could easily pass through the patient’s pelvic inlet plane (Figure 1C). Overall, these CT pelvimetry findings suggested no CPD.
Pregnancy outcomeThe patient wished to avoid labor induction. Labor started naturally at 3 weeks after CT pelvimetry. Fetal BPD (measured by ultrasound) increased from 87 to 92 mm (+5.8%), while the estimated fetal weight increased from 2,316 to 2,650 g (+14.4%). We re-analyzed the CT images of the maternal pelvis and confirmed the absence of CPD.
At a gestational age of 39 weeks and 1 day, the patient vaginally delivered a healthy male baby (2,732 g). The infant’s 1- and 5-min Apgar scores were 8 and 9, respectively, and arterial cord blood pH was 7.34. The mother and baby were discharged on postpartum day 5 without morbidity.
Construction of 3D models using a 3D printerAfter anonymization, CT data were sent to a 3D printing team at Hotty Polymer Co. (Tokyo, Japan), where 3D models were constructed using a Raise3D Pro3 3D printer. Polylactic acid resin was used for the pelvic bone model, and thermoplastic elastomers (TPE Soft Touch –60 Shore A) were used for the fetal head bone model. The latter has some plasticity that allows for emulation of the shape change during labor. A +10% magnified fetal head model was also constructed.
Labor simulationBy the time we received the 3D models, the patient had already delivered her baby vaginally; thus, the following simulations were performed for research purposes only. Using the 3D pelvis and fetal head models, we simulated a trial of labor with the following criteria established for diagnosing disproportion: (1) Easy: the fetal head model passed through the pelvic canal without pushing; (2) Difficult: the fetal head model passed through the pelvic canal by pushing or manual rotation; and (3) Impossible: the fetal head model was unable to pass through the pelvic canal despite pushing or manual rotation. Further simulations were performed, including (1) abnormal internal rotation, such as the occiput posterior, and (2) fetal head model diameter magnified by +10%. To our knowledge, no previous report has described the situations examined in these simulations. We added the second set of simulations to confirm that our models can detect CPD.
Figure 1D shows a 3D image of the maternal pelvis and fetal head. Using CT data, 3D models were constructed with a 3D printer (Figures 2A and 2B) and a trial of labor was simulated. Three experts evaluated the simulation results, and all determined that the fetal head could easily pass through the pelvic canal inlet (Figure 3A), followed by the pelvic canal outlet (Figure 3B), without pushing. The fetal head could pass through the pelvic canal even in the posterior occiput position (Figure 4A). Using the fetal head model magnified by 10%, all three judges determined that the fetal head could not pass through the narrow part of the mid-pelvis despite pushing or manual rotation (Figure 4B).
CT-based pelvimetry with lower fetal radiation exposure emerged as a substitute for X-ray pelvimetry, which was once considered the standard method for diagnosing CPD. Our low-dose CT system achieves a substantially lower exposure dose (0.17 mGy, 10 mAs) than that of conventional X-ray pelvimetry (1.18 mGy), or even compared with the dose reported in previous low-dose CT studies (0.39 mGy).18) We performed CT phantom simulations at 10 mA to minimize fetal radiation exposure, as sufficient diagnostic image quality was obtained at both 10 and 20 mA. The obtained images enabled us to accurately measure the obstetric diameters of the maternal pelvis and fetal head. Our findings support the hypothesis that low-dose CT pelvimetry is superior to conventional x-ray pelvimetry.
The use of low-dose CT for diagnosing CPD requires balancing its potential benefits and risks. The primary benefit is that an early CPD diagnosis can significantly reduce maternal and fetal morbidities associated with labor obstruction. An early and accurate diagnosis allows physicians to recommend a planned cesarean section and avoid emergency procedures, prolonged labor, and associated complications such as uterine rupture, fetal distress, and birth-related trauma. These benefits, however, must be weighed against the risk of fetal radiation exposure. Although our low-dose CT protocol delivers less radiation (0.17 mGy) than conventional X-ray pelvimetry, fetal radiation exposure carries a theoretical risk. While the current scientific consensus suggests that fetal radiation doses <50 mGy are not associated with increased congenital malformation or pregnancy loss rates, the ALARA principle still applies.11) Thus, constructing 3D models based on non-radiative diagnostic methods should be considered in the future.
In 1991, Abitbol et al. reported that CPD should be suspected when the cephalopelvic disproportion index is <9 mm.17) Shimaoka et al. reported a correlation between APDPO and emergency cesarean delivery in 78 cases.16) To date, however, no systematic review has presented a single CPD marker with high sensitivity and specificity for predicting labor outcomes.9) In 2007, Rozenberg strongly criticized pelvimetry, noting that well-designed randomized studies are lacking to support the utility of X-ray pelvimetry for predicting CPD.19) Indeed, only one study has reported that pelvimetry had no influence on neonatal issues or increased caesarean section rates,19) highlighting the need for a new approach. Our current findings suggest that 3D model-based labor simulations may be a useful strategy, providing all necessary information on the maternal pelvic bone as well as the fetal head. As the present study was preliminary in nature, using data from only one patient, the efficacy of our models must be validated through further investigation.
The primary advantage of 3D model-based labor simulation is that it can be used to model any scenario. For example, we were able to easily try fetal head rotations and malrotations, and to create several sizes of 3D fetal head models to mimic fetal head growth. This time, we constructed a +10% magnified fetal head model and confirmed that it could not pass through the pelvic canal. It should be noted that +10% is a theoretical value used to test our idea.
This study has several limitations. First, our simulations did not completely imitate the plasticity of the fetal skull during labor. We used a somewhat plastic material to create the fetal head model; however, improvements are required to mimic the mechanism of fetal head molding during labor. Second, our 3D bone models were not designed to directly evaluate the soft birth canal tissue. Rather, the original CT images provided information regarding these structures. Therefore, future studies should explore the incorporation of soft tissue analysis from CT data into the simulation process. Third, uterine contraction strength cannot be fully accounted for using static models. Thus, among the five components of CPD stated by Mengert20) –1) bony pelvis size and shape, 2) fetal head size, 3) uterine force exerted, 4) fetal head moldability, and 5) fetal presentation and position – our simulation system offers good insight into components 1, 2, and 5. Although our 3D modeling technique offers promising improvements in CPD diagnostics, it does not eliminate all uncertainty.
In conclusion, our approach offers several advantages. First of all, pelvic and fetal data are visualized as palpable models, allowing patients to physically interact with the models of their own pelvis and fetal head and even perform simulations. Moreover, simulation results are comprehensible, making it possible for medical staff and patients to easily share objective clinical data. These features will facilitate shared decision-making.
We thank Editage (www.editage.jp) for English language editing.
None.
