Israel Medical Association Journal

Artificial intelligence (AI) has become the emblem of progress. We are told it learns faster, sees patterns invisible to the human eye, and will soon outthink us in every domain, from finance to philosophy, from language to life. In medicine, where decisions carry the weight of saving lives, this narrative has gained traction. Algorithms promise precision without fatigue, accuracy without bias, and reproducibility without emotion. Yet, sometimes, the data tell a quieter story.

Artificial intelligence (AI) and machine learning have moved to the forefront of scientific discourse and clinical medicine, offering improved accuracy and efficiency while raising concerns about transparency, accountability, and unintended consequences. Recent developments, particularly large-scale and generative models, have fueled these debates. However, efforts to mimic aspects of human intelligence long predate ChatGPT. These efforts include the early rule-based systems to Weizenbaum’s ELIZA program, which humorously simulated a Rogerian psychotherapist in its Doctor script [1]. For clinicians, the real test is not whether predictions become marginally more accurate on average, but whether they improve the identification of high-risk patients and meaningfully change management.

Background: The sonographic assessment of estimated fetal weight (EFW) is essential for identification of fetuses in weight extremes and aids in peripartum management. However, there are inconsistent reports regarding EFW accuracy.

Objective: To examine maternal and fetal determinants associated with unreliable EFW.

Methods: A retrospective case-control study was conducted at a single, tertiary medical center between 2011 and 2019. All term, singleton deliveries with a sonographic EFW within 2 weeks of delivery were included. Unreliable EFW was defined as > 500 grams discordance between it and the actual birth weight. We allocated the study cohort into two groups: unreliable EFW (cases) and accurate EFW (controls).

Results: Overall, 41,261 deliveries met inclusion criteria. Of these, 1721 (4.17%) had unreliable EFW. The factors positively associated with unreliable EFW included body mass index > 30 kg/m², weight gain > 20 kg, higher amniotic fluid index, pregestational diabetes, gestational age > 41^0/7, and birth weight ≥ 4000 grams. On multiple regression analysis, pregestational diabetes (odds ratio [OR] 2.22, 95% confidence interval [95%CI] 1.56–3.17, P < 0.001) and a higher birth weight (OR 1.91, 95%CI 1.79–2.04, P < 0.001) were independently associated with unreliable EFW. On analysis of different weight categories, pregestational diabetes was associated with unreliable EFW only among birth weights ≥ 3500 grams (OR 3.28, 95%CI 1.98–5.44, P< 0.001) and ≥ 4000 grams (OR 4.27, 95%CI 2.31–7.90, P < 0.001).

Conclusion: Pregestational diabetes and increased birth weight are independent risk factors for unreliable EFW and should be considered when planning delivery management.

Background: Sonographic estimation of birth weight may differ among evaluators due to its operator-dependent nature.

Objectives: To compare the accuracy of estimation of fetal birth weight by sonography between ultrasound-certified physicians and registered diagnostic medical technicians.

Methods: The authors reviewed ultrasound examinations that had been performed by either technicians or ultrasound-certified obstetricians between 2010 and 2017, and within 2 days of delivery. Inclusion criteria were: singleton viable pregnancy, details of four ultrasound measurements (abdominal circumference, bi-parietal diameter, head circumference, and femur length), and known birth weight. The estimated fetal weight (EFW) was calculated according to the Hadlock formula, incorporating the four ultrasound measurements. The mean percentage error (MPE) was calculated by the formula: (EFW-birth weight) x100 / birth weight.

Results: Technicians performed 9741examinations and physicians performed 352 examinations. The proportion of macrosomic neonates was similar in both groups. Technicians were more accurate than physicians in terms of the MPE, absolute MPE, proportion of estimates that fell within ± 10% of birth weight, and Euclidean distance (P < 0.0001 for all comparisons). They were also more accurate in terms of sensitivity, specificity, positive predictive value, negative predictive value, and area under the receiver operating curve. Furthermore, for fetuses weighing more than 4000 grams the technicians had a lower total false prediction rate.

Conclusions: Medical technicians in our institute performed better than physicians in estimating fetal weight. Further studies are warranted to confirm our findings and better delineate the role of repeat physician’s examination after an initial estimation by an experienced technician.

Background: Two types of growth curves are commonly used to diagnose fetal growth disorders: neonatal birth weight (BW) and sonographic estimated fetal weight (EFW). The debate as to which growth curve to use is universal.

Objectives: To establish sonographic EFW growth curves for the Israeli population and to assess whether the use of the BW growth curves currently adapted in Israel leads to under-diagnosis of intrauterine growth disorders.

Methods: Biometric data collected during a 6 year period was analyzed to establish sonographic EFW growth curves between 15–42 weeks of gestation for the Israeli population. Growth curves were compared to previously published sonographic EFW growth curves. A comparison with the Israeli BW growth curves was performed to assess the possibility of under-diagnosis of intrauterine growth disorders.

Results: Out of 42,778 sonographic EFW studies, 31,559 met the inclusion criteria. The sonographic EFW growth curves from the current study resembled the EFW curves previously published. The comparison of the current sonographic EFW and BW growth curves revealed under-diagnosis of intrauterine growth disorders during the preterm period. Four percent of the fetuses assessed between 26–34 weeks would have been suspected of being growth restricted; 2.8 percent of the fetuses assessed between 30–36 weeks would have been suspected of having macrosomia, based on the BW growth curves.

Conclusions: New Israeli sonographic EFW growth curves resemble previously published sonographic EFW curves. Using BW growth curves may lead to the under-diagnosis of growth disorders. We recommend adopting sonographic EFW growth to diagnose intrauterine growth disorders.