Correlation of Prenatal Diagnosis and Pathology Findings Following Dilation and Evacuation for Fetal Anomalies Open Access

We conducted a retrospective cohort study analyzing the clinicopathologic correlation of fetal anomaly cases after surgical pregnancy termination in the second trimester. The Committee on Human Research at the University of California, San Francisco (UCSF), reviewed and approved this study.

Our study is based on 448 pregnancies diagnosed with fetal anomalies by prenatal imaging studies and/or prior cytogenetic analysis, resulting in therapeutic abortions by D&E. All procedures were done at Zuckerberg San Francisco General Hospital and Trauma Center (ZSFG), a hospital affiliated with UCSF. Dilation and evacuation procedures are done at ZSFG up to 24 weeks of gestation. Most patients undergo 1 day of cervical preparation with osmotic dilators, although some patients presenting in the early second trimester undergo same-day procedures. We searched our pathology records for nonintact fetal specimens obtained via D&E between January 2008 and December 2013, and we included specimens between 12 and 24 weeks of gestation dated by ultrasound examinations performed immediately before D&E procedures. We excluded all cases in which D&E followed intrauterine fetal demise. Additionally excluded were structural anomaly cases that lacked a complete ultrasound report and chromosomal aberration cases that lacked detailed laboratory karyotype reports. We also omitted all cases with skeletal anomalies in which specimen radiographs had not been obtained as part of the pathology examination.

In addition to gross examination of fetal and placental tissues, we also performed histologic examinations, usually by preparing 3 or more microscopic slides. We consistently performed histologic examination of organs or tissues for which microscopic findings were either necessary or helpful for determining the final diagnoses. To confirm fetal sex, gonads were sampled in all cases in which they were identified. As a rule, at least 1 microscopic slide containing placenta, umbilical cord, and extraplacental membranes was examined. Typically, representative microscopic sections of normal-appearing organs were also evaluated.

We created a database of clinical and pathology findings by using FileMakerPro (FileMaker, Inc, Santa Clara, California). Initially, we categorized fetal anomaly cases into 1 of 2 groups (termination based on abnormal ultrasound findings versus abnormal genetic testing results). We categorized the clinicopathologic correlation between ultrasound and pathology diagnosis as complete agreement, partial agreement (confirmation of at least 1 defect detected on ultrasound examination), diagnostic discrepancy, or extensive fragmentation preventing adequate examination. For pregnancies terminated for abnormal karyotypes, we also determined how frequently anomaly patterns representative of specific chromosomal aberrations were found on pathology examination. Criteria for evaluation in cases with abnormal genetic testing results were based on pathology findings and were classified as follows: pathology findings consistent with the chromosomal abnormality, no abnormal findings present, or fragmentation of organ/body site preventing adequate evaluation.

Complete ultrasound reports were also available in most of the included chromosomal abnormality cases. We combined these cases with those terminated for structural defects and performed a comprehensive correlation of ultrasound findings with pathology diagnoses obtained from pathology reports. In addition to final diagnoses, pathology reports included a detailed description of gross anatomic and histopathologic findings. To correlate with our findings on pathology examination, each ultrasound finding was categorized into 1 of the following 9 anomaly groups: central nervous system (CNS), orofacial clefts, cystic hygroma/nuchal edema, cardiac, body wall/diaphragm, gastrointestinal, genitourinary, musculoskeletal, or miscellaneous abnormalities. The data were analyzed by using Analysis ToolPak add-in for Excel Office 365 (Microsoft Corporation, Redmond, Washington). The t test was used when appropriate. A P value of less than .05 was considered to be statistically significant.

From a total of 549 D&E cases, we identified 448 that fulfilled our criteria and were included in this study. We excluded cases for the following reasons: intrauterine fetal demise (73 cases), incomplete ultrasound records (17 cases), unobtainable karyotype reports (4 cases), or lack of specimen radiographs (7 cases).

The mean age (±standard deviation) of women undergoing D&E owing to structural anomalies detected by imaging studies was 28.2 ± 6.5 years; the mean age of women with fetuses diagnosed with an abnormal karyotype was 34.6 ± 7.1 years (P < .001). Based on ultrasonography examination immediately before termination, the mean gestational duration for fetuses with structural abnormalities was 20.3 ± 2.7 weeks, compared to 19.2 ± 3.1 weeks for fetuses terminated for chromosomal aberrations (P < .001). Most fetal anomaly cases (278 of 448 fetuses, 62%) were terminated owing to structural congenital abnormalities. Abnormal karyotypes led to termination of pregnancy in the remaining 170 cases (38%).

The 170 chromosomal abnormalities included trisomy 21 (92 cases, 54%), trisomy 18 (36 cases, 21%), trisomy 13 (13 cases, 8%), and others (29 cases, 17%) (Table 1). Structural anomalies were identified by pathology examination in 109 of the 170 chromosomal aberration specimens (64%). These findings were diverse but typically included features associated with the specific disorders. Of the remaining 61 cases, 28 (17%) did not have anatomic defects and 33 (19%) were deemed too fragmented to assess adequately.

Trisomy 21 was the most frequent chromosomal abnormality resulting in termination of pregnancy, making up 92 of the 170 chromosomal abnormality cases (54%). Of these, 22 specimens (24%) were extensively fragmented and could not be evaluated for findings characteristic of this anomaly. However, 53 fetuses with trisomy 21 (58%) showed at least 1 structural abnormality on pathology examination and a total of 80 pathology findings were noted. Of these, 64 (80%) were characteristic findings, 7 (9%) were infrequently observed, and 9 (11%) were not usually identified in fetuses with trisomy 21. Frequently associated abnormalities included an atrial septal defect, ventricular septal defects, atrioventricular septal defects, dysplastic ears, flat nose, sandal gap, intrauterine growth retardation, and single transverse palmar creases. In 17 specimens, structural abnormalities were not identified in the tissues and organs available for study.

Genetic testing revealed trisomy 18 in 36 of the 170 fetuses with chromosomal abnormalities (21%). Pathology examination established the presence of at least 1 structural abnormality in 32 of the 36 (89%). In total, 72 abnormal findings were observed in the 36 fetuses with trisomy 18. Of these, 51 defects (71%) were those that are generally found in individuals affected by trisomy 18 and included clenched hands/overlapping fingers, rocker-bottom feet, dysplastic ears, low-set ears, growth deficiency, micrognathia, single umbilical artery, and ventricular septal defects. Other defects were findings less commonly associated with this karyotype. Fragmentation prevented recognition of structural changes in 3 cases. One case did not show structural abnormalities on ultrasound examination, and this was confirmed by thorough pathologic examination of a minimally fragmented specimen.

Thirteen pregnancies were terminated after a fetal diagnosis of Patau syndrome. Characteristic anomalies were present in all except 1 severely fragmented specimen. Pathologists noted 30 structural abnormalities among these specimens, including commonly seen trisomy 13 abnormalities (cleft lip, cleft palate, clenched hands, polydactyly and posterior prominence of the heel, and malformed ears). The remaining but less frequently encountered anomalies were clubfoot deformity, cyclopia, edema, flattened nose, hypotelorism, proboscis, sandal gap, and single uterine artery.

Twenty-nine additional cases included various chromosomal abnormalities. Of the 16 cases (55%) in which structural abnormalities are part of the phenotype, characteristic findings were identified in 8 (50%). Prader-Willi/Angelman syndrome and a β-globin mutation are not associated with structural fetal anomalies and none were detected. The remaining 11 cases (38%) represent rare chromosomal abnormalities in which associated structural abnormalities are poorly characterized. Of these, genetic testing revealed 6 deletion syndromes. Specimens diagnosed with the karyotypes 46,XX,del(1)(q41q43); 46,XY,del(2)(q23q31); and 46,XY,del(1)(p36.3p36.3)(pter−,CDK118−) showed anomalies on pathology examination. A fetus with the karyotype 46,XX,del(5)(p16.1) showed no structural abnormalities. Fragmentation of fetal tissue prevented adequate evaluation in 2 deletion syndrome specimens, namely, 46,XY,del(10)(p12) and 46,XX,del(22)(q13). In the 5 cases with rare karyotypes, anomalies were found in 46,XY,der(8)t(8;8)(p23.1;q21.2). Karyotypes 46,XY,der(6)t(6;21)(q25.3;p11.2)mat and 45,XY,der(18;21)(q10;q10) showed no structural anomalies. Evaluation was hindered by extensive fragmentation in specimens with the karyotypes 46,XX,add(3)(p25) and 47,XX,t(4;10)(p16;p11.2),+mar.

Ultrasound reports were available for review in 413 fetal anomaly cases—278 cases (67%) terminated owing to structural defects detected on ultrasound examination and 135 cases (33%) terminated owing to chromosomal abnormalities. Correlations with pathology diagnoses for each group are displayed in Table 2. Among the 278 congenital structural anomaly cases diagnosed with prenatal ultrasonography, complete clinicopathologic agreement was seen in 69 cases (25%), partial agreement in 97 (35%), whereas the clinical and pathologic diagnosis differed in 11 cases (4%). In 10 cases (∼4%), the defects identified on ultrasonography could not be confirmed owing to disruption of the relevant anatomic site; however, pathology examination did yield abnormal findings in other organs or body sites. Extensive fragmentation prevented adequate pathologic examination in 91 cases (32%). One hundred thirty-five cases with abnormal prenatal karyotypes (79%) had prenatal ultrasound examinations. Of these, there was complete clinicopathologic agreement seen in 30 cases (22%) and partial agreement in 36 cases (27%), whereas the clinical and pathological diagnosis differed in 19 cases (14%). In 12 cases (9%) disruption of relevant anatomic site prevented correlation, but again, pathology examination revealed abnormal findings in other sites. Extensive fragmentation prevented adequate pathologic examination in 38 of the 135 chromosomal abnormality cases with abnormal prenatal ultrasound findings (28%).

A total of 738 specific congenital defects were noted in the prenatal ultrasound studies from both groups. Table 3 summarizes the occurrence, distribution, and correlation of findings among our 9 categories of congenital anomalies, and these are discussed below.

Overall, 89 pregnancies (21.5%) were terminated on the basis of the ultrasound diagnoses of CNS abnormalities alone. A total of 214 congenital CNS anomalies, including 28 among the cases with chromosomal aberrations (13%), were identified on prenatal ultrasonography and were studied (Supplemental Table 1; see supplemental digital content containing 8 tables at www.archivesofpathology.org in the February 2017 table of contents). In most CNS defects (183 findings, 86%), correlation was prevented by fragmentation of brain tissue and spinal cord structures. Extensive fragmentation of intracranial contents prevented evaluation of all 110 intracerebral abnormalities found on prenatal imaging studies. Clinicopathologic agreement between ultrasound and pathology findings was only seen with 31 defects (14%), all of which were neural tube defects. Of these, a prenatal diagnosis of acrania/anencephaly was confirmed in 22 of 35 instances (61%), making it the congenital CNS anomaly most likely to be preserved after D&E. A single case of holoprosencephaly, not noted on the prenatal ultrasound report, was identified during pathology examination; however, this case was unusual in that cervical dilation allowed the extraction of an intact specimen. One case of spina bifida not identified by prenatal screening procedures was detected on pathology examination.

Four diagnoses of orofacial clefts were cited as the reason for termination. Of these, 3 (75%) ultimately were confirmed, while in 1 (25%) a pathology diagnosis could not be obtained owing to tissue disruption. Overall, ultrasound examinations of 413 fetuses resulted in the identification of 20 orofacial clefts (5%), and the condition of 10 (2.5%) of these specimens allowed correlation (Supplemental Table 2). In 5 cases of cleft lip (25%), 1 case of cleft palate (5%), and 1 case of combined cleft lip and palate (5%), pathologists were able to verify prenatal ultrasound diagnosis. Two fetuses with orofacial clefts were also diagnosed with amniotic band syndrome. In one, a prenatal finding of cleft lip was confirmed. In the other with a prenatal diagnosis of combined cleft lip/cleft palate, fragmentation prevented verification. Three orofacial clefts (15%) (2 cleft lip; 1 cleft lip/palate) were erroneously diagnosed on prenatal screening. Notably, pathology examination revealed 14 orofacial clefts that were not identified on ultrasound examination. Of these, 11 cases (79%) were associated with other structural congenital anomalies detected prenatally; 3 (21%) were noted in fetuses with abnormal karyotypes.

We identified 38 cases of suspected cystic hygroma (9%) among the 413 fetuses with available ultrasound reports. Fetal nuchal tissue was severely fragmented and could not be assessed in all but 1 confirmed case (97%). Nuchal translucency—an indication of fetal nuchal edema—was observed on ultrasonography in 30 specimens (7%). Pathology examination of nuchal structures was prevented owing to severe fragmentation in all but 1 case (97%), in which the anatomy of the neck was well preserved and appeared normal on examination.

Ninety cardiac defects were diagnosed on ultrasound studies (Supplemental Table 3). In 25 of the 90 cardiac findings (28%), ultrasound diagnosis was substantiated by the pathology examination. Hypoplastic left heart syndrome was the cardiac anomaly most reliably confirmed. Nineteen heart defects (21%) previously identified on ultrasonography could not be confirmed despite thorough examination of cardiac structures including the great vessels. Extensive disruption of cardiac structures prevented evaluation and subsequent correlation of 46 findings (51%), including 12 atrioventricular septal defects (13%) and 12 ventricular septal defects (13%). Overall, 24 additional cardiac defects not seen on prenatal imaging studies were noted on pathology examination. Our study includes 17 fetuses terminated solely after prenatal diagnoses of congenital heart disease. Complete agreement was seen in 11 cases (65%). Discrepancies were seen in 2 fetuses (12%) in which the abnormal cardiac structures described on ultrasonography were normal on pathology examination. Cardiac structures were extensively fragmented in another 4 cases (24%). Routine ultrasound screening resulting in a diagnosis of a cardiac anomaly had prompted fetal echocardiograms in 9 cases (53%). The echocardiographic diagnoses confirmed the preliminary diagnoses in 8 cases (89%) and slightly modified the prenatal diagnosis in 1 (11%). In this subgroup, there were no diagnostic discrepancies identified in 5 intact hearts available for subsequent pathology study.

Prenatal ultrasound diagnosis of body wall or diaphragm defects led to termination of 9 pregnancies. Thirty-six body defects involving the body wall or diaphragm had been noted on prenatal ultrasound studies (Supplemental Table 4). Adequate evaluation was only possible for 5 of these (14%). Agreement was seen in 2 fetuses with omphaloceles, 1 with congenital diaphragmatic hernia, and 1 with gastroschisis. In 1 fetus, an abdominal wall abnormality subsequently recognized on pathology examination as prune belly syndrome was identified on the ultrasound study, though its exact nature was unspecified.

In our series of cases, termination of pregnancy was never initiated for anomalies affecting the gastrointestinal tract alone. Twenty-eight abnormalities affecting the gastrointestinal system were visualized on ultrasonography (Supplemental Table 5). Pathologists confirmed the diagnoses for 9 (32%); however, abnormalities were absent in 16 fetuses (57%) (1 case of presumed tracheoesophageal fistula [3%], 7 cases of suspected gastric agenesis [25%], and 8 cases of hyperechogenic bowel [29%]). In 3 specimens (11%), ultrasound diagnosis could not be confirmed owing to extensive fragmentation of the gastrointestinal tract. Of note, pathology examination revealed 5 gastrointestinal malformations not detected during prenatal imaging studies, including 2 cases with imperforate anus and 1 with rectal atresia, all of which showed other anomalies consistent with the VACTERL association (vertebral anomalies, anal atresia, cardiac defect, tracheoesopegeal fistula, renal anomaly, and limb anomalies).

Termination was based exclusively on congenital genitourinary anomalies in 14 cases and the clinical diagnosis was verified in all (Supplemental Table 6). Of 62 congenital genitourinary defects identified on ultrasonography, pathology examination was able to confirm 43 (69%). Clinicopathologic correlation for congenital obstructive renal disorders was 100% (6 findings), followed by multicystic renal dysplasia with 92% (11 of 12 findings). Pathology examination revealed a false-positive prenatal ultrasound diagnosis for 6 presumed abnormalities (10%), 5 of which were suspected agenesis of genitourinary structures. Disruption of the urinary tract prevented evaluation in 13 specimens (21%). A total of 12 findings (19%), including 4 cases of multicystic renal dysplasia (6.3%), 4 cases of horseshoe kidneys (6.3%), and 4 cases consisting of other genitourinary anomalies (6.3%), were not identified on ultrasound studies but were subsequently discovered on pathology examination.

Thirteen fetuses had anomalies limited to the musculoskeletal system, with clinicopathologic agreement of the major findings in 10 of these cases (77%) (Supplemental Table 7). A total of 108 congenital musculoskeletal anomalies were identified during prenatal ultrasound studies. Sixty of these 108 findings (55%) were seen as part of an anomaly constellation involving multiple organ systems. Pathology examination confirmed 59 (55%) of these. Seventeen abnormalities (16%) identified by ultrasonography were not confirmed. Extensive fragmentation prevented adequate evaluation of 32 ultrasound findings (30%), including anomalies of the skull and facial structures, extremities, spine, and sternum. Pathology examination identified 72 additional defects, of which 60 (83%) affected hands or feet.

Defects affecting organs and body system other than the anomaly groups discussed above were categorized in a separate group of miscellaneous findings. We assigned 112 congenital anomalies to this category (Supplemental Table 8). Clinicopathologic agreement was found with 56 abnormalities (50%). Ultrasound findings were not confirmed for 22 abnormalities (20%). Severe disruption of organs or body sites of interest hindered correlation for 34 findings (30%). However, 108 additional defects were noted on pathology examination.

Standard D&E procedures are the most frequent method of elective abortion in the second trimester.^12–14  The resulting pathology specimens usually receive a cursory surgical pathology examination primarily to document the presence of chorionic villi and fetal tissues. Despite its potential for preventing accurate pathology diagnoses, therapeutic D&E procedures are replacing labor inductions for fetal anomalies and intrauterine fetal deaths.^5,6,15  Although a detailed, systematic examination of D&E specimens obtained after intrauterine fetal demise, or for anomalies diagnosed prenatally, can confirm, clarify, extend, or perhaps contradict the prenatal diagnostic studies, there is a paucity of studies evaluating the utility of pathology examination of fragmented specimens.^7,11  Establishing the utility and limitations of pathologic examinations of D&E specimens is important, since accurate final diagnoses of congenital anomalies—in intact fetuses—have been shown to alter subsequent genetic counseling.¹⁶  Thus, in our study, we correlated prenatal diagnoses of congenital anomalies with pathologic findings in D&E specimens, specifically comparing various organ systems or body regions and particular chromosomal defects. In 413 cases with prenatal ultrasound findings, there was complete agreement in 99 cases (24%), partial agreement in 133 cases (32%), and disagreement in 30 cases (7%). Clinical diagnoses could not be confirmed, owing to disruption of the relevant organs and/or body sites in 22 cases (5%), but were supplemented by findings detected on pathology examination of other anatomic sites. Extensive fragmentation completely prevented evaluation in the remaining 129 cases (31%). In studies of intact specimens, complete clinicopathologic agreement ranged from 47% to 91%.^17–23  In a smaller but detailed study of 36 D&E specimens, Sun et al¹⁰  found complete agreement in 30% of cases, comparable to our results.

We also determined in how many cases pathology examination supplemented or altered prenatal ultrasound diagnoses, finding that pathology examination yielded at least 1 additional finding that augmented the prenatal diagnoses in 137 cases (33%). This is higher than the percentage reported by Vogt et al²²  (16%) but lower than that reported by Rajashekar et al²⁴  (60%) in their studies of fetal and perinatal autopsies. Pathology examination corrected prenatal ultrasound diagnoses in 175 cases (37%), a rate approximately 10% lower than that of Klatt⁸  who in a small series of 37 D&E cases found that pathology examination was required to reach a specific diagnosis in 17 (46%) cases. We did not determine whether our findings altered the estimated recurrence risk. However, Boyd et al,²⁵  in an autopsy study of 132 induced terminations carried out for ultrasound-diagnosed congenital anomalies, found that pathology examination added information in 35 cases (27%), reducing recurrence risk in 17 (13%) and increasing it in 18 (14%).

The greatest degree of correlation was seen in congenital genitourinary anomalies, in which pathology examination confirmed 43 of the 62 findings (69%) affecting the kidneys, ureters, bladder, and urethra, comparable to the clinicopathologic correlation of 73% in an autopsy study of 308 intact, second-trimester fetuses.²⁶  Severe genitourinary defects such as renal agenesis (11 cases), obstructive renal dysplasia (6 cases), and posterior urethral valves (4 cases) were reliably confirmed on pathology examination of D&E specimens in 82% (9 cases), 100%, and 100% of cases, respectively. Eleven of the 12 multicystic kidneys (92%) identified on ultrasonography were confirmed on gross pathology examination, though microscopy was often required to classify these as either cystic or obstructive renal dysplasias. Pathologists discovered 4 multicystic renal diseases as a supplemental finding, suggesting that this diagnosis can be challenging to recognize by ultrasound examination. All 4 cases of horseshoe kidneys were described on pathology study only, indicating that the isthmus is difficult for sonographers to recognize. Cho et al²⁷  advocate measuring the renal pelvic angle to improve prenatal detection of this malformation.

Of 90 congenital cardiac anomalies detected on ultrasonography, 25 of our findings (28%) were confirmed, the same correlation rate found by Sun et al¹⁰  in their study of D&E specimens, but much lower than the overall 74% to 82% correlation rate of congenital heart defects reported in studies of intact infants and fetuses.^26,28  Although heart tissue was generally identified in our D&E specimens, sufficiently intact specimens were only available in about half the cases, similar to what has been described previously.⁷ 

One hundred eight abnormalities affecting the musculoskeletal system were described on prenatal ultrasonography. Pathology examination substantiated the clinical diagnosis in 59 findings (55%), approximately 5% lower than the clinicopathologic correlation of prenatally diagnosed musculoskeletal defects following examination of intact fetuses.¹⁷  Fragmentation of relevant body sites prevented correlation in 32 of 108 findings (30%). The fragmentation rate of 30% (32 findings) in the musculoskeletal system group was the second lowest of our 9 groups. This group also showed a particularly high number of supplemental findings not detected on prenatal ultrasound examinations owing primarily to difficulties in recognizing subtle abnormalities affecting the hands or feet, a known limitation of prenatal ultrasonography.²⁹ 

Gastrointestinal malformations represented a small fraction of fetal anomalies detected in the prenatal studies. Twenty-eight gastrointestinal anomalies were described in fetuses that had either multiple other anomalies (15 findings, 54%) or chromosomal aberrations (13 findings, 46%). Disruption of fetal tissue hindered adequate evaluation in 3 of these findings (11%), which was the lowest fragmentation rate among all groups. Notably, in our cases, there was a 57% false-positive rate for gastrointestinal anomalies detected on ultrasound examinations. Pathology examination did not confirm 16 findings, including 7 cases with a diagnosis of gastric atresia and 8 descriptions of echogenic bowel. However, echogenic bowel is considered a sonographic normal variant or soft marker, and unless associated with intrauterine growth retardation, additional structural anomalies, or chromosomal anomalies, the ultimate prognosis is favorable.^30,31 

Evaluating CNS, nuchal area, and body cavity/diaphragm anomalies proved to be most challenging because these body sites are particularly prone to tissue disruption.⁷  However, confirmation of anencephaly was an exception, as this diagnosis could be confirmed on pathology examination in most suspected cases. Consistent with the results of previous studies,^7,10  neural tube abnormalities were the CNS defects most likely to be preserved.

Anomalies confined to the nuchal area proved very difficult to evaluate in our D&E specimens. Nuchal tissue was severely fragmented in all but 2 of the 68 specimens (97%) in which cystic hygroma or nuchal edema was seen on prenatal ultrasonography. This is in stark contrast to intact specimens, where prenatally diagnosed cystic hygroma and nuchal edema could be confirmed by pathologists in 93% and 72% of cases, respectively.²⁹  Ernst et al⁷  recommend a D2-40 stain as a supplemental tool for diagnosis of cystic hygroma in D&E specimens; however, this technique was not used in any of our cases.

In summary, this study was undertaken to evaluate the effectiveness of pathologic examination after termination of pregnancies by D&E during the second trimester. Pathology findings corrected prenatal ultrasound diagnoses in 152 of the 413 cases (37%). Taking into account all of our cases, pathology examinations detected 237 findings not described on prenatal ultrasound studies, yielding additional diagnostic information in 137 of 413 cases (33%). Therefore, our study supports the capability of pathology examinations in confirming, correcting, or refining prenatal diagnoses even in fragmented fetal tissues obtained after D&E. Our data lend further validation for performing detailed pathology examinations of these specimens in order to evaluate congenital anomalies and provide the most complete data possible for subsequent genetic counseling. Pathology examination of fetuses with rare chromosomal anomalies or poorly characterized syndromes can provide additional phenotypic data. Furthermore, pathology diagnoses afford opportunities for quality control of ultrasound studies that may subsequently improve their diagnostic accuracy. The ability of pathologists trained in the evaluation of fragmented fetal specimens may reassure genetic counselors and women preferring D&E that they do not necessarily need to have induction termination for the purpose of obtaining intact fetuses for accurate pathologic examinations. However, our findings reaffirm the difficulty of evaluating most abnormalities involving the brain and nuchal region, and defects affecting the body wall or diaphragm, following D&E procedures. Thus, if complete pathology examination is desired, induction termination should be recommended.