Article Text
PDF
Abstract
Aims Specific identification of a hydatidiform mole (HM) and subclassification of a complete hydatidiform mole (CHM) or partial hydatidiform mole (PHM) are critical. This study aimed to reappraise the diagnostic performance of ultrasonography and histology with a refined diagnosis.
Methods This was a retrospective, multicentre cohort study of 821 patients with histologically suspected HM specimens. Refined diagnostic algorithms with p57 immunohistochemistry and short tandem repeat (STR) genotyping were performed and used as the true standard for assessing the diagnostic performance of the original ultrasonography and morphology methods. The diagnostic performance was calculated using accuracy, agreement rate, sensitivity and the positive predictive value (PPV) compared with refined diagnostic results.
Results Of the 821 histologically suspected HM cases included, 788 (95.98%) were successfully reclassified into 448 CHMs, 213 PHMs and 127 non-molar (NM) abortuses. Ultrasonography showed an overall accuracy of 44.38%, with a sensitivity of 44.33% for CHM and 37.5% for PHM. The overall classification accuracy of the original morphological diagnosis was 65.97%. After exclusion of the initially untyped HMs, the overall agreement rate was 59.11% (κ=0.364, p<0.0001) between the original and refined diagnoses, with a sensitivity of 40.09% and PPV of 96.05% for diagnosing CHMs and a sensitivity of 84.98% and a PPV of 45.59% for diagnosing PHMs. The interinstitutional variability of morphology in diagnosing HMs was significant among the 15 centres (range, 8.33%–100.00%, p<0.0001).
Conclusion The current diagnosis of HM based solely on ultrasound or morphology remains problematic, and ancillary techniques, particularly p57 immunohistochemistry and DNA genotyping, should be integrated into routine practice as much as possible.
- histopathology
- gestational trophoblastic disease
- diagnosis
- pathology, molecular
- placenta
Data availability statement
Data are available on reasonable request. All data relevant to the study are included in the article or uploaded as supplementary information. Not applicable.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Accurate identification and subclassification of hydatidiform moles are essential, with both ultrasound and histopathological assessments being subject to observer variability.
The diagnostic efficacy of ultrasonography and morphological examination in the context of hydatidiform mole diagnosis has not been thoroughly evaluated in large, multicentre cohorts.
WHAT THIS STUDY ADDS
We conducted a retrospective cohort analysis to assess the diagnostic performance of ultrasonography and morphology in identifying hydatidiform moles, employing a refined diagnostic approach with p57 immunohistochemistry and molecular genotyping.
Subsequently, we analysed and compared the interinstitutional variability observed in histological diagnoses of hydatidiform moles, using data compiled from multiple centres.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
Re-evaluation of the initial ultrasonographic and histopathological findings revealed areas for improvement in diagnostic accuracy.
It is recommended that ancillary diagnostic techniques be increasingly integrated into routine clinical practice to enhance the reliability and consistency of hydatidiform mole diagnosis.
Introduction
Hydatidiform mole (HM) is a type of genotypically abnormal gestation, affecting approximately 1/1000 pregnancies.1 Specific identification and subclassification of HMs are critical because a molar pregnancy is non-viable and carries risks of recurrence and malignant transformation in the future. Approximately 15%–20% of complete hydatidiform moles (CHMs) and <5% of partial hydatidiform moles (PHMs) progress to gestational trophoblastic neoplasia (GTN), necessitating surgery and/or chemotherapy2; conversely, persistent GTN following non-molar gestation is rare (≤0.0002%).3 Consequently, patients with HMs are followed up with serial serum β-human chorionic gonadotropin (β-hCG) measurements in conjunction with contraception.
Ultrasound and histopathological examination are the main diagnostic methods currently used, are necessary for the diagnosis of HM and have a major impact on patient management. However, both ultrasound and histopathology are observer dependent and not highly accurate due to the overlapping manifestations of molar and non-molar (NM) pregnancies.4 The sensitivity of ultrasound for accurate detection of HM is reported to be <50%,5 6 with moderate inter-observer agreement (κ=0.41).7 Pathologically, according to a previous study approximately 30% of CHMs and 71% of PHMs are misdiagnosed.8 Even among pathology experts, significant inter-observer and intra-observer variability have been observed.9 10 Therefore, obtaining a precise diagnosis of HM based solely on ultrasonography or histology remains challenging.
Several ancillary techniques have been developed and recommended for integration into clinical practice to refine the diagnosis of HM. Among them, p57 immunohistochemistry analysis and DNA genotyping of short tandem repeat (STR) loci have demonstrated their value in numerous studies.1 11 p57 is encoded by the paternally imprinted gene CDKN1C and can reliably distinguish CHM from PHM and NM pregnancies with a sensitivity of 98.4% and specificity of 62.5%.12 However, PHMs cannot be distinguished from NMs using p57, so DNA genotyping is required to differentiate these two entities. DNA genotyping is currently considered to be the ‘diagnostic truth’ in the confirmation and subtyping of sporadic HMs because of its ability to trace the parental origin of chromosome complements.13 14 Considering the high cost of genotyping and the excellent performance of p57 immunohistochemistry, a diagnostic algorithm approach combining the two above-mentioned methods has been recommended to refine the diagnosis of HM.1 11 15 16 Xing et al prospectively analysed 2217 patients at a single institution using this diagnostic algorithm.17 However, studies that have previously used large cohorts tended to be prospective analyses that focused on screening missed cases instead of overdiagnosed cases.17 18
This study aimed to examine cases where HM was suspected based on histopathology to calculate the accuracy, sensitivity and positive predictive value (PPV) of ultrasonography and histology. To do this, we performed a retrospective cohort study to explore the performance of ultrasonography and histology in the diagnosis of HM via a refined diagnosis with p57 immunohistochemistry and molecular genotyping. Then, we compared the interinstitutional variability of histology in diagnosing HM based on data obtained from the multiple centres.
Materials and methods
Study design and cases
This was a retrospective and multicentre cohort study of 924 formalin-fixed paraffin-embedded (FFPE) products of conception (POC) intrauterine specimens retrieved from patients with histologically suspected HMs including consultation cases, in-house cases and various collaborators’ pathology laboratories in Zhejiang Province, China, from March 2019 to July 2023.
Specifically, those who met any of the following criteria were included in this study: (1) the initial morphological diagnosis of HMs was made solely through H&E staining by the surgical institution’s pathology report; these results served as the ‘original morphological diagnosis’ and were used to assess its accuracy in subsequent evaluations below; (2) a supplementary diagnosis of HMs was established through ancillary molecular techniques, including p57 immunohistochemistry, DNA genotyping or single-nucleotide polymorphism microarray analysis and (3) cases with an original non-molar diagnosis were reconsidered for HM after undergoing pathological review at other hospitals. The exclusion criteria were as follows: (1) inability to obtain any pathological reports from the surgical institution; (2) absence of villous tissue or presence of only isolated villous clusters or (3) H&E staining of POC samples indicating the presence of a severe intermixture of maternal and villous tissues that could not be adequately separated even by microdissection, preventing subsequent molecular analysis.
The clinical information, including maternal age at the time of pregnancy, gestational age, serum β-hCG levels, ultrasound reports, original histopathological reports and as much of the pregnancy history as possible, was obtained by a survey sent to collaborators’ laboratories or patients. See online supplemental table 1 for more detailed cases information.
Supplemental material
According to the recognised refined diagnostic algorithms for HMs,1 15 specimens were centrally diagnosed in our laboratory via H&E staining, p57 immunohistochemical staining and genotyping. A group of four gynaecological pathologists (XZ, BH, XY and HC) was assembled to evaluate all the slides and interpret the molecular diagnosis results, culminating in a consensus-based definitive refined diagnosis. All pathologists were blinded to the original pathology reports from the surgery institutions. Given the excellent performance of the p57 assay, for p57-negative patients, CHM diagnosis was established; for p57-positive patients, genotyping was performed. In addition, we conducted STR genotyping analysis for the following cases even if they had a negative p57 status: (1) morphology equivocal: defined as very early pregnancy specimens exhibiting atypical features or those lacking an internal positive control despite negative p57 status; (2) adverse pregnancy history: defined as patients with recurrent spontaneous abortion or HMs and (3) patients with a consultation request: defined as patients who proactively sought consultation willingly joined the study, and provided samples for additional genotyping tests.
p57 immunohistochemistry
A mouse monoclonal antibody against the p57 protein (MXB Biotechnologies, Fuzhou, China) was used for p57 expression immunohistochemical analysis. The p57 immunostaining results were interpreted as follows:
Negative: villous cytotrophoblasts and stromal cells had either no expression or limited expression of p57 (nuclear staining in <10%), with intermediate trophoblastic cells and/or maternal decidual cells that expressed nuclear p57 used as the internal positive control.
Positive: p57 staining was extensive or diffuse (expression >50%) in all cells or focally positive (>10% but <50%) in both villous cytotrophoblasts and stromal cells.
Divergent: there were two populations of villi, each with different morphologies and staining patterns.
Discordant: positive staining in cytotrophoblast but negative staining in villous stromal cells, or vice versa (no such cases were involved in this study).
Unsatisfactory: the staining result could not be interpreted, most commonly for reasons such as loss of immunoreactivity, technical failure or lack of an internal positive control.
STR genotyping
Five to 10 serial sections, each 5 μm thick, were generated from the POC FFPE blocks of each patient. Deparaffinised and hydrated sections were H&E stained without coverslips as we previously reported.19 Pure chorionic villi and maternal tissues were scraped separately by microscopic dissection. PCR-based STR analysis of 15 STR loci was performed under thermal cycling conditions, and capillary electrophoresis was carried out according to the manufacturer’s instructions (AmpFlSTR Identifiler Plus PCR Amplification Kit, Foster City, California, USA) on an ABI3130 platform. Data collection and analysis were performed using Gene Marker software V.3.7 (Biogene, Cambridge, UK).
Allelic ratios were calculated by the peak height to make a diagnosis according to a previous study.20 The molecular genotyping results were interpreted as follows:
CHM: the presence of purely paternal allele patterns without maternal chromosome complements(s).
PHM: the presence of a paternal:maternal allele ratio of 2:1 at all loci, indicating diandric triploidy, with rare triandric tetraploids not involved in this study.
NM: the presence of a balanced biparental genetic profile, including trisomy, which showed isolated copy changes involving only one or two STR loci.
Unsatisfactory: an amplification result that cannot be interpreted, most commonly for reasons such as insufficient villi, maternal contamination, or unsuccessful PCR amplification.
Statistical analysis
Analysis of variance and Tukey’s post hoc test were used for multiple-group comparisons of data. The refined diagnostic results with p57 immunohistochemistry and STR genotyping were used as the gold standard (true) diagnosis for assessing diagnostic performance and calculating the accuracy rate. The diagnostic agreement was tested using kappa statistics. Associations among categorical variables were tested using the χ2 test or Fisher’s exact test, depending on the sample size. Statistical significance was set at p<0.05. Statistical analyses were performed using SPSS statistical software (V.19; IBM, Armonk, New York, USA).
Results
Refined diagnostic results
Among the 924 total POC specimens collected, pathology evaluations and interpretation of the results were conducted by a consensus group of four pathologists (XZ, BH, XY and HC). In total, the specimens of 821 of 924 patients (88.85%) were processed for p57 immunohistochemical analysis, and 523 were subjected to genotyping (figure 1).
Diagnostic flow chart and summary characterisation of products of conception results tested by the refined diagnostic algorithm with p57 immunohistochemistry and short tandem repeat genotyping. BiCHM, biparental complete hydatidiform mole; CHM, complete hydatidiform mole; CHMCF, twin pregnancy with a CHM and a coexisting fetus; FFPE, formalin-fixed paraffin-embedded; HM, hydatidiform mole; PHM, partial hydatidiform mole.
In all, 803 of 821 specimens (97.81%) had satisfactory p57 immunohistochemical analysis results, with 18 (2.19%) having non-reactive or suboptimal expression results. Of these, 445 specimens that were negative for p57 were diagnosed with CHMs (figure 2), and the remaining 355 patients were p57 positive and evaluated by STR genotyping. In three cases of twin pregnancy, the components displayed a divergent p57 staining pattern characterised by coexisting p57-negative and p57-positive villi, and the diagnosis of twin pregnancy with a complete HM and a coexisting fetus (CHMCF) was further confirmed by molecular genotyping (figure 3).
Representative morphological characteristics and genotyping results of androgenetic complete hydatidiform mole (CHM). (A) A typical CHM shows enlarged hydropic villi with circumferential trophoblastic hyperplasia, cistern formation, trophoblastic inclusion and apoptotic bodies within the villous stroma. (B) Negative p57 expression in villous cytotrophoblasts and stromal cells (intermediate trophoblastic cells were positive and served as an internal positive control). (C) Genotyping demonstrated purely androgenetic patterns of the products of conception (POC, bottom panel) with only a novel/paternal allele without a maternal allele (upper panel).
Representative morphological characteristics and genotyping results of twin pregnancies with a complete hydatidiform mole and coexisting fetus (CHMCF). (A) CHMCF has two distinct villi populations: markedly enlarged oedematous villi coexist with small normal villi. (B) Divergent p57 expression: p57-negative molar component and p57-positive non-molar component present in the same specimen. (C) Genotyping demonstrated that the normal villous tissue contained a maternal allele and a paternal allele, with an allele ratio of 1:1 indicating biparental conception (bottom panel), but the molar tissue (middle panel) inherited only the paternal allele, indicating purely androgenetic conception.
In total, 523 patients, including 165 p57-negative patients, 355 p57-positive patients, and 3 p57-divergent patients, underwent molecular genotyping. Satisfactory results were not achieved for 15 patients due to unsuccessful amplification. Of the 165 p57-negative patients, 160 were confirmed to have androgenetic CHMs, and 5 had biparental CHMs (BiCHMs). This reaffirmed the outstanding performance of the p57 assay for diagnosing CHM. All the patients who provided these five BiCHM POC specimens had two or more HMs, and one of them was confirmed to be a carrier of a homozygous nonsense mutation in the NLRP7 gene, as we reported previously.21 Among the 355 p57-positive patients, 213 (60.00%) had diandric triploidy, which was confirmed as PHM (figure 4), and 127 (35.77%) were diagnosed with non-molar abortuses with balanced biparental alleles and hydropic changes and/or abnormal villous morphology (figure 5).
Representative morphological characteristics and genotyping results of a partial hydatidiform mole. (A) Classical enlarged fibrotic to hydropic and irregularly scalloped villi with moderate trophoblastic hyperplasia and some trophoblastic inclusions. (B) Positive p57 expression in villous cytotrophoblasts and stromal cells. (C) Genotyping demonstrated digenic triploidy of the products of conception (POC) (bottom panel) with a paternal: maternal allele ratio of 2:1 at all informative loci.
Representative morphological characteristics and genotyping results of non-molar abortuses. (A) Non-molar conceptions have some mild hydropic changes and some variable irregular shapes with polarised trophoblastic hyperplasia. (B) Positive p57 expression in villous cytotrophoblasts and stromal cells. (C) Genotyping demonstrated a balanced biparental pattern of the products of conception (POC) (bottom panel) with a paternal:maternal allele ratio of 1:1.
Overall, 788 patients obtained a definitive classification by refined diagnosis with p57 immunohistochemistry and molecular genotyping, resulting in the diagnosis of 448 CHMs, 213 PHMs and 127 NM abortuses.
Clinical characteristics
We obtained 627 provider surveys with a completion level of >50% for analysis of additional clinical information (table 1 and online supplemental table 1). All the patients had intrauterine pregnancies, and there were no egg donors. The mean maternal age at the time of pregnancy loss for all diagnostic categories was similar (29–32 years). As previously reported,22 CHMs in this series of cases showed a bimodal distribution, peaking in maternal age groups under 21 and over 45 years (figure 6A). The average gestational age ranged from 62 to 78 days (8.85–11.14 weeks) among the three groups. Patients with HMs had excessive hCG levels; in particular, hCG levels >100 000 mIU/mL were detected in 52% (127/241) of patients with CHMs and 40.91% (36/88) of patients with PHMs, consistent with previous reports.23 Nine of 490 women (1.84%) reported a previous molar pregnancy, and they were all confirmed to have a recurrent molar pregnancy with a diagnosis of CHM. Among the 262 patients who were successfully followed up in this study, 40 (24.69%) with CHMs and 1 (1.41%) with PHM developed GTN, consistent with previous data.24
Clinical data on maternal age distribution and the relationship between gestational age and ultrasound detection rate. (A) Percentage of patients with CHM, PHM and NM in stratified age groups. (B) Gestational age distribution at the time of the evacuation of products of conception in each group. (C) The ultrasonic detection rate in each group corresponding to gestational age (weeks). CHM, complete hydatidiform mole; HM, hydatidiform mole; NM, non-molar pregnancy; PHM, partial hydatidiform mole.
Clinical characteristics of 627 patients who had refined diagnostic tests and provided their information
Performance of the original ultrasound diagnosis regarding hydatidiform moles
A total of 169 additional ultrasound reports were obtained in the provider survey, and all were included in the study. As presented in table 2 and online supplemental table 2, diagnoses were refined, resulting in 97 diagnoses of CHMs, 56 diagnoses of PHMs and 16 diagnoses of NMs. Overall, the classification accuracy of ultrasound was 44.38% (75/169), with a sensitivity of 44.33% (43/97) for CHM, 37.5% (21/56) for PHM and 68.75% (11/16) for NM. There was a slight trend towards an increasing ultrasound detection rate with increasing gestational age, which was related mainly to CHM rather than PHM (figure 6B,C). From the ultrasound reports, we identified ‘heterogeneous masses’ and ‘snowstorms’ as the most common terms in CHM diagnostic summaries, with sensitivities of 60.82% and 38.14%. In PHM reports, ‘no cardiac activity’, ‘snowstorm’ and ‘heterogeneous mass’ were prevalent, with sensitivities of 41.07% and 30.36%.
Supplemental material
The original ultrasound data and findings in 169 patients
Performance of the original morphological diagnosis
A total of 788 patients obtained a definitive refined diagnosis, and 21 (2.66%) patients had negative p57 immunostaining results integrated into their initial morphological reports. To avoid bias, we excluded these patients from this and the next sections. Therefore, the comparison of the original histopathological diagnosis with the refined diagnostic results was conducted for 767 patients who underwent definitive reappraisal. Among them, 136 (17.73%) were initially diagnosed with untyped HM and then reclassified as CHM (n=129), PHM (n=4) or NM (n=3; table 3). The overall classification accuracy of the original morphological diagnosis was 65.97% (506/767). After these untyped patients were excluded (n=631), the agreement rate was 59.11% (373/631), with a kappa value of 0.364 (p<0.0001), indicating poor diagnostic consistency (figure 7A). For the diagnosis of CHM, the sensitivity was 40.09%, the PPV was 96.05% and the respective values for PHM were 84.98% and 45.59%. The most common problem in the misclassification of cases by morphology was the distinction of PHMs. Of the 397 initially subtyped PHMs, a total of 115 (28.97%) were subsequently confirmed to be CHMs, and 100 (25.19%) were confirmed to be NMs. Due to confusion regarding the divergent pathology and p57 staining pattern, two CHMCFs were misdiagnosed: one as PHM and another as singleton CHM.
The original histological diagnostic performance. (A) The agreement rate between the original morphological diagnosis and the refined diagnosis, excluding the initially untyped hydatidiform mole (HM) cases (n=631). (B) The correct classification rate of each institution for diagnosing HMs and partial hydatidiform moles (PHMs). CHM, complete hydatidiform mole.
Performance of the original morphological diagnosis
Interinstitutional variability of histology
Since the diagnosis of HMs based solely on morphology is affected by inter-observer and intra-observer variability, we investigated whether interinstitutional variability exists. After excluding institutions providing fewer than 10 cases, 15 institutions with 646 definitively reclassified cases were included for further analysis. For more details about these institutions, see online supplemental table 3. The overall diagnostic accuracy of the original pathologically correct classification of each centre ranged from 8.33% to 100.00% (overall 63.00%), with significant differences among them (χ2 test, p<0.0001) (table 4, and figure 7B). Because of this imbalance in the data, more patients (351/646, 54.33%) who were initially pathologically diagnosed with PHMs were included, and we observed institutional variability in the accuracy of the morphological diagnosis of PHMs. The diagnostic accuracy of the PHMs ranged from 0.00% to 100% (overall 49.00%) (table 4 and figure 7B). When the HMs that were initially untyped according to histology were excluded, the κ value of each centre ranged from 0.468 (poor) to 1.00 (perfect) (p<0.0001), indicating that there was significant interinstitutional variability in the histological diagnosis of HMs.
Supplemental material
Centre-specific diagnostic performance of morphology for HMs
Discussion
In this study, 788 cases of initially suspected HMs according to histopathology were retrospectively confirmed via p57 immunohistochemistry and STR genotyping, and the results demonstrated that the diagnostic agreement rate between the initial and refined diagnoses was only 44.38% according to ultrasonography and 65.97% according to histology, with significant interinstitutional variability. As indicated in previous studies, ultrasound or morphology-based methods are limited, and molecular ancillary techniques combined with genetic testing are warranted.25
The literature on the diagnosis of HMs is replete with studies demonstrating poor reproducibility,5 7 9 26 and our study also confirmed this; notably, distinguishing PHMs from NMs is the main challenge. For ultrasonic diagnosis, we reported a similar sensitivity (40%–50%) and detection rate (30%–50%) of HMs to the values reported in previously published studies.27–29 Until ultrasound detection improves considerably, routine histological evaluation of the POC will be necessary to exclude HMs.
Regarding the morphological diagnosis of HMs, numerous prospective studies have been designed to determine the incidence and missed diagnosis rate of HMs or to verify the applicability of auxiliary diagnostic techniques.10 15 30 In this retrospective study, we identified problems in the histological misdiagnosis and overdiagnosis of molar pregnancy. Sporadic studies have reported that the original pathological diagnosis was misclassified in 45% of cases,31 similar to our findings. Notably, 16.12% of suspected HM cases were reclassified to NM pregnancy in our study, which means that these patients underwent unnecessary clinical follow-up despite no increased risk of GTN. Given the excellent performance of p57 immunostaining, the diagnosis of CHM is no longer a challenge. As we reported above, the PPV of PHMs on histology was only 45.59%, and the main problem was the frequent need to differentiate PHMs from NMs. Besides, we first identified the interinstitutional variability of HM by histology. Multiple variables may lead to variation between institutions: (1) some patients may not have developed the typical morphological features, especially in the first trimester, during which detection is related to the institutional experience with ultrasonic; (2) histological criteria for HM are not being consistently applied among pathologists26; (3) the grade of hospital or institution affects the number of patients and the quality of education, thus affecting the experience and level of pathologists; (4) institutional variation in tissue sampling strategies for POC may also result in different villi distributions for evaluation31 and (5) an imbalance in the number of cases among different institutions causes bias in the data analysis. Therefore, we encourage cross-institutional consultation for suspicious cases and the integration of ancillary techniques into routine practice.
Two of the most commonly applied ancillary techniques are p57 immunohistochemistry and STR genotyping, but they also have several limitations. Although p57 is a particular and sensitive marker of CHM, misdiagnosis is possible, false-negative p57 expression can occur under the following conditions: (1) loss of maternal chromosome 11 in non-CHM cases18 32; (2) mutation of the maternal allele of the p57 gene, such as in Beckwith-Wiedemann syndrome in NM pregnancy33 and (3) paternal uniparental disomy of chromosome 11.34 Some atypical staining patterns are interpreted as ‘discordant’ or ‘divergent’ p57 immunostaining. However, p57 expression is highly correlated with genotyping results in general; only 0.6% of the patients in the previous study and 0.38% (3/788) of the patients in the current study had aberrant p57 expression,17 making it the only universally accepted immunohistochemical marker of HMs. Therefore, given satisfactory p57-negative results, molecular genotyping is generally not necessary for routine diagnosis of CHMs.
Compared with other gene test techniques, including karyotyping, DNA ploidy analysis and fluorescence in situ hybridisation, STR genotyping has the advantage of tracing the parental origin of chromosome complements.35 The limitations of STR genotyping include its relatively high cost and challenges involving the interpretability of its results in unique contexts, such as mosaic conceptions,36 37 BiCHMs38 and maternal contamination.39 However, the diagnostic algorithm we adopted can provide a warning regarding these situations from morphology and p57 immunostaining evaluation, which largely avoids misdiagnosis. In fact, in our study, 35.77% of the patients with p57-positive immunostaining triggering genotyping were confirmed as NM pregnancies initially overdiagnosed as HMs, further emphasising the necessity of genotyping.
The main limitations of our study were its retrospective nature and the fact that the cohort was based on consultation services or voluntary declarations of cases, which is different from population-based epidemiological studies. Furthermore, multi-institutional or hospital-based studies have limitations because they are subject to patient referral bias. However, this multicentre study comprised a wide spectrum of data, suggesting that the database is not unduly biased.
Conclusion
In conclusion, a series of 788 histologically suspected HMs were reclassified by refined diagnosis with the modern approach of p57 immunohistochemistry and STR genotyping. Reappraisal of the original ultrasonography and histology results showed imperfect performance, with accuracies of 44.38% and 65.97%, respectively. Significant variability was also observed among different institutions. Given the high misdiagnosis rate here, we recommend integrating ancillary techniques into routine practice as much as possible.
Data availability statement
Data are available on reasonable request. All data relevant to the study are included in the article or uploaded as supplementary information. Not applicable.
Ethics statements
Patient consent for publication
Ethics approval
The institutional review boards approved this study, and the requirement for informed consent was waived because of its retrospective nature (institutional review board no. 2020IIT339).
Acknowledgments
We acknowledge and thank the patients and the collaborators for their participation in this study.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Handling editor Sarah Chiang.
YZ and LC contributed equally.
Contributors JQ and XZhang were responsible for the study concept and design. YZ and LMC designed and performed the experimental work. YZ and BH analysed the data. XZhang, BH and XY assisted with the pathological review and interpretation of results. XY, DP, JD, HC, JL, HS, ZZ, LJ, XZhu and XZhang contributed to sample collection. All authors critically reviewed and edited the manuscript and approved the final version, ensuring the integrity of the work. JQ is the guarantor.
Funding This work was supported by the National Key Research and Development Programme of China (No. 2023YFC2705801), the Key Research and Development Programme of Zhejiang Province (No. 2020C03116), Ningbo Public Welfare Science and Technology Project (2021S156) and the National Natural Science Foundation of China (No. 82071665 and 82371669).
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.