Update on the pathology of gestational trophoblastic disease€¦ · Gestational trophoblastic...

Review Article

Update on the pathology of gestational trophoblastic

disease

DEBRA S. HELLER

Department of Pathology & Laboratory Medicine, Rutgers-New Jersey Medical School, Newark, NJ, USA

Heller DS. Update on the pathology of gestational trophoblastic disease. APMIS 2018; 126: 647–654.

Gestational trophoblastic disease can be a challenging area for pathologic evaluation. An update with a focus onpathologic challenges is presented.

Key words: Gestational trophoblastic disease; hydatidiform mole; choriocarcinoma; trophoblastic neoplasms.

Debra S. Heller, Department of Pathology & Laboratory Medicine, Rutgers New Jersey Medical School, 185 SouthOrange Ave-UH E158, Newark, NJ 07103, USA. e-mail: [email protected]

Gestational trophoblastic disease (GTD) means dif-ferent things to the pathologist and the clinician. Thepathologist utilizes a classification based onhistopathology, while the clinician uses a clinicallybased classification, and does not always need tosubmit tissue for histopathology to appropriatelymanage the patient, particularly with post-molarGTD (Table 1). Persistent trophoblastic disease aftera mole may be a persistent mole, an invasive mole(Fig. 1), or much more rarely, choriocarcinoma(Fig. 2). Management may not require a histopatho-logic distinction. There are good in-depth reviews ofboth pathologic and clinical aspects of GTD (1–3).This review focuses on updates and challenges forthe pathologist. Challenges include distinction ofmoles from non-molar gestations, recognition ofearly moles, distinction of complete moles from par-tial moles, familiarity with lesions of extravillous tro-phoblast (EVT), and recognition of forms of GTDthat may be seen by the pathologist after therapy.

HYDATIDIFORM MOLE

Hydatidiform moles are divided into complete andpartial hydatidiform moles. These are geneticallydifferent lesions, with complete moles usually com-prised of a diploid complement, both paternallyderived, while partial moles are diandric triploids in

most cases. Digynic triploids are not moles. Rarely,complete moles may be biparental, in familial com-plete moles, associated with mutations of theNLRP7 gene (4). Genetic variations from the usualchromosomal complements seen in molar diseasehave been reported. Recognizing moles and distin-guishing complete from partial moles is importantdue to differing risk of persistent trophoblastic dis-ease, which is considerably greater after a completemole (15–20%) than a partial mole (<5%). Recur-rence rates of moles in future pregnancies are thesame (1–2%) (2). Histology can be challenging.Below are some specific issues that can arise.

Early complete mole

Hydatidiform mole was previously diagnosed mostoften in the second trimester, when marked uterineenlargement, passage of ‘grapes’, bleeding, and otherclinical manifestations became apparent. With theadvent of early ultrasounds in pregnancy, most com-plete moles are detected in the first trimester, beforefull development of the pronounced histologic fea-tures seen in later complete moles, which show theexpected edematous avascular villi with marked cir-cumferential trophoblast proliferation. Earlier com-plete moles have distinct features as well, andfamiliarity with these features will aid in detection.Features include redundant bulbous terminal villi,hypercellular myxoid stroma, labyrinthine villousstromal capillaries, focal atypical villous andReceived 10 May 2017. Accepted 24 September 2017

647

APMIS 126: 647–654 © 2018 APMIS. Published by John Wiley & Sons Ltd.

DOI 10.1111/apm.12786

JOURNAL OF PATHOLOGY, MICROBIOLOGY AND IMMUNOLOGY

extravillous trophoblast hyperplasia, focal extravil-lous trophoblast hyperplasia, hyperchromatic exag-gerated implantation site trophoblast, and stromalkaryorrhexis and apoptosis (Fig. 3) (5, 6)

Distinguishing complete from partial moles and moles

from non-molar gestations

Even with subspecialty expertise, diagnosis of molesby evaluation of hematoxylin and eosin stained slidesis challenging, with poor interobserver reproducibility(7). Partial moles have less pronounced histopatho-logic features than complete moles, with less edemaand trophoblast proliferation. The presence of fetaltissue, including nucleated red blood cells or fetalmembranes, favors a partial mole over a completemole, but rare complete moles have evidence of fetaltissue, including a complete mole with a twin, or amosaic complete mole (6). In partial moles, villi tendto be irregular and scalloped, with trophoblast inclu-sions, which can also be seen in pregnancies withother chromosomal abnormalities, although thosepregnancies will not show trophoblast proliferation(Figs 4 and 5). This differs from the more roundedvilli of complete moles. Missed abortions can alsoshow edematous avascular villi, although the atypicaltrophoblastic proliferation of complete moles isabsent. In order to further characterize these lesions, avariety of adjunctive techniques have been employed.

A B

Fig. 3. Early complete hydatidiform mole showing bulbous projections from villi (A), myxoid stroma with labyrinthinevessels and proliferating extravillous trophoblast (B).

Fig. 1. Invasive mole-Molar villi with adjacent necrosiswithin myometrium.

Fig. 2. Choriocarcinoma-A biphasic proliferation of cellsis seen in the absence of chorionic villi.

Table 1. Classification of gestational trophoblasticneoplasia

Pathologic classificationHydatidiform mole-complete/partialInvasive moleChoriocarcinomaPlacental site trophoblastic tumor (PSTT)Epithelioid trophoblastic tumor (ETT)Miscellaneous Trophoblastic Lesions-Exaggeratedplacental site, Placental site nodule/plaqueUnclassified

Clinical classificationBenign Gestational trophoblastic Disease

Hydatidiform molesMalignant Gestational Trophoblastic Disease

Non-metastaticMetastatic w/good prognosisMetastatic w/poor prognosis

Reference: Shih et al. (25).

648 © 2018 APMIS. Published by John Wiley & Sons Ltd

HELLER

P57 immunohistochemistry

P57 immunostaining is helpful, but it is necessary toknow how to interpret the results, which are not ‘pos-itive’ and ‘negative’. P57 is a paternally imprintedmaternally expressed gene which stains maternal

tissues. This allows decidua to serve as an internalpositive control. In complete moles, there is no villousstaining, although extravillous trophoblast clustersdo stain (Fig. 6A). The explanation for this stainingof extravillous trophoblast clusters is unclear, but hasbeen theorized to be due to incomplete or relaxedimprinting of p57, or less likely due to cross-reactivity(8). In partial moles, there is staining of villous stro-mal cells and cytotrophoblast (Fig. 6B).

Other adjunctive techniques

Ploidy analysis, via flow cytometry, fluorescentin situ hybridization or chromogenic in situhybridization, can be helpful, distinguishingdiploidy from triploidy. However, ploidy will notdistinguish a diploid missed abortion from a diploidcomplete mole, or a triploid digynic gestation froma triploid diandric partial mole.

Short tandem repeat (STR) genotyping hasemerged as the most useful ancillary technique, notonly distinguishing moles from their mimics, andcomplete mole from partial mole, but also able todistinguish gestational from non-gestational tro-phoblastic disease, as well as recurrent from newmolar disease (7, 9). STR genotyping can be per-formed on formalin-fixed, paraffin-embedded tissue.The limiting factor is the lack of availability of thetechnique at all centers. As in all techniques, thereare rare unusual cases and pitfalls.

Familial hydatidiform mole (biparental complete

mole)

A rare clinical form of complete hydatidiform moleis familial recurrent hydatidiform mole, where themoles are diploid but biparental. This is a recurrentautosomal recessive disorder. Histologically, thelesions are similar to sporadic complete moles, andusually are histologically indistinguishable, but themolar features may be less pronounced. P57immunostaining is the same as for sporadic com-plete moles (4).

Fig. 4. Partial mole-Villi are scalloped, with less tro-phoblast proliferation than complete mole, however it iscircumferential. There is a mix of molar and non-molarvilli. Evidence of fetal tissue (not shown) should be sought.

Fig. 5. Trophoblast inclusions-Numerous trophoblastinclusions are seen in this triploid placenta.

A B

Fig. 6. P57 immunostaining in complete (A) and partial (B) moles. In complete moles (A), there is staining of extravilloustrophoblast, but villi are negative. In partial moles (B), there is staining of villous stromal cells and cytotrophoblast.

© 2018 APMIS. Published by John Wiley & Sons Ltd 649

GESTATIONAL TROPHOBLASTIC DISEASE

MIMICS OF HYDATIDIFORM MOLE AND

GTD

A variety of lesions may be mistaken for hydatidi-form mole or GTD. Features of these are illus-trated below:

Placental mesenchymal dysplasia

Placental mesenchymal dysplasia is a rare lesionassociated with chromosomal abnormalities, Beck-with-Wiedemann syndrome, and stillbirth. It mani-fests as placentomegaly with cystic dilatation ofstem villi which can be mistaken for molar vesicles,as opposed to molar dilatation of terminal villi (10)(Fig. 7). The cystic dilatation is concentrated nearthe fetal surface of the placenta, giving it a uniquegross appearance.

Trophoblast inclusions

Partial moles are characterized by scalloping of ter-minal villi. These indentations can be seen on histo-logic sections as trophoblast inclusions if sectionedtangentially. While a rare one is probably of no sig-nificance, if increased over 1–2, trophoblast inclu-sions can also be a manifestation of otherchromosomal abnormalities, including triploidy (11)(Fig. 5), and increased trophoblast inclusions arenot a rare finding in spontaneous abortions. Theseother abnormal gestations will lack the trophoblastproliferation and mix of molar and non-molar villiof a partial mole.

Chorangioma with trophoblastic proliferation

Chorangiomas are fairly frequent in placentas, andrepresent hemangiomas within the parenchyma.

Unless large, they are usually of no clinical signifi-cance. Some of these lesions are associated withtrophoblast proliferation around the periphery, ortracking along lobules of the lesion (Fig. 8), andthe literature has termed these lesions ‘chorangio-carcinomas’. This is an unfortunate misnomer, asthere have been no reports of persistent trophoblas-tic disease in mothers or their infants (12).

Florid trophoblastic proliferation in early pregnancy

and tubal ectopic pregnancy

An area of challenge is the exuberant trophoblastproliferation that can be seen in early intrauterinegestations, as well as in tubal ectopic pregnancy.The trophoblast proliferation in early non-molargestations has been described as ‘polar capping’,with the trophoblast at one end of the villus, theside of the direction of implantation. It is possiblefor a tubal hydatidiform mole to occur, but it isexceptionally rare. P57 Immunostaining may help ifthis issue arises in the fallopian tube or earlyintrauterine gestation. If products of conception areexamined early enough for there to be no chorionicvilli (pre-villous trophoblast), the question of chori-ocarcinoma may arise. In the absence of availabilityof STR genotyping, the distinction may fall on clin-ical evaluation, however this is a rare issue.

INTRAPLACENTAL CHORIOCARCINOMA

Intraplacental choriocarcinoma is not seen often,and this may be due to the lack of a grossly recog-nizable lesion in some cases. It is rare, representingabout 0.03% of GTD in one series (13). It is gener-ally found on examination of third trimester

Fig. 7. Placental mesenchymal dysplasia-Stem villi areenlarged with a myxoid stroma and prominent vessels.

Fig. 8. Chorangioma with trophoblast proliferation-Tro-phoblast is seen in the center between lobules of a choran-gioma.


HELLER

placentas. While usually asymptomatic, vaginalbleeding, melena as well as respiratory and neuro-logic symptoms have been reported (13). Histologi-cally, there is marked trophoblast proliferationaround one or a cluster of villi, while the rest ofthe placenta is normal (Fig. 9). This is an impor-tant lesion to recognize and report to the clinicians,as there may be metastatic disease in both mothersand infants (13). Intraplacental choriocarcinomamay be the precursor lesion of gestational chorio-carcinoma after a term gestation.

PITFALLS IN FOLLOWING SERUM BETA-

HCG

Beta-hCG testing can give false positive resultsunrelated to GTD. This can occur in some tumors(14), end-stage renal disease, with some drugs, withheterophile antibody (15), or by misinterpretationof results. A variety of conditions can give falsepositive urine hcg results (16). False negative results

have been reported with GTD (17), and clinicaljudgment is always important.

A form of hCG, hyperglycosylated hcg (hCG-H)is made by extravillous cytotrophoblasts, asopposed to usual hCG, which is produced by syn-cytiotrophoblasts. hCG-H is related to normalimplantation, and is associated with aggressivenessas well as chemosensitivity of choriocarcinoma,where it tends to be higher (18). It is lower in lessaggressive/more chemoresistant GTD.

Quiescent GTD is a presumed inactive form ofGTD with low levels of true hCG over severalmonths, with disease undetectable, and the hCGunresponsive to chemotherapy. hCG subanalysisreveals no hyperglycosylated hCG (18). Long-termlow b-hCG with no detectible disease may revert tonormal at some point, but in one series (19), a largeproportion of non-reverters were subsequently dis-covered to have PSTT or epithelioid trophoblastictumor (ETT).

POST-EVACUATION GTD WITH FACTORS

COMPLICATING HISTOPATHOLOGIC

DIAGNOSIS

Tissue received by the pathology laboratory for eval-uation of persistent GTD after mole can be challeng-ing, and this can be complicated by factors such asinsufficient history, or prior therapy. Curettings mayonly contain scant individual trophoblast cells, some-times atypical, in blood clot. Theoretically, that find-ing could represent persistent mole, invasive mole,choriocarcinoma, PSTT/ETT, or a new intrauterinegestation. Lu et al. (20) described four cases of GTDwhere masses persisted after chemotherapy, twomoles, one ‘GTD’, and one choriocarcinoma. Afterchemotherapy, the tissue resembled ETT, and theauthors suggested that chemotherapy may have elim-inated the more primitive choriocarcinoma cells,allowing evolution into a more chemoresistant phe-notype of extravillous trophoblast.

Fig. 9. Intraplacental choriocarcinoma-Exuberant tro-phoblast proliferation around a cluster of villi in a thirdtrimester placenta that was otherwise normal. Necrosis isnot uncommon.

A B

Fig. 10. Persistent GTD – A nodule is seen in the lower uterine segment (arrow) (A). It was mainly comprised of fibrinand blood clot, with rare atypical trophoblast cells dispersed within (B).



A lack of history can also hamper diagnosis. Ahysterectomy for GTD was received by the labora-tory for a persistent mass (Fig. 10A). Histologicevaluation showed rare atypical cells within bloodclot (Fig. 10B). The initial history was that thepatient had had a mole evacuated somewhere else afew months prior, and there was something ‘funnyabout it’. The patient’s serum hCG was reportedlylow. Based on the greater immunostaining withhuman placental lactogen than with beta-hCG ofthe atypical cells from the hysterectomy, the initialimpression was of a possible lesion of extravilloustrophoblast, such as PSTT. Several weeks later, theprior slides from the molar pregnancy were avail-able for review. They showed a complete mole witha more than usual amount of trophoblast prolifera-tion, with greater atypia. The original diagnosissuggested that this mole may have been evolvinginto choriocarcinoma, and that the criteria for thiswere not well established. The reclassified finaldiagnosis on the hysterectomy was ‘GTD’, nolonger thought to represent a lesion of extravilloustrophoblast. This case emphasizes the importanceof review of prior contributory pathologic material.

LESIONS OF EXTRAVILLOUS

TROPHOBLAST

There are three types of EVT, the EVT of the tro-phoblastic columns, implantation site EVT, andchorionic EVT. These different types of EVT havedifferent immunohistochemical profiles, as do thelesions arising from them (Table 2) (21). This cate-gory of lesion can be diagnostically challenging, dueto overlap of histologic features, as well as lack offamiliarity. EVT has also been termed ‘intermediatetrophoblast’ in the literature, and some authors con-tinue to utilize this terminology, but this is basedmore on histopathologic features than on the cellsrepresenting an intermediate stage of development.

Placental site nodules/plaques

Placental site nodules and plaques are frequentlyincidental findings in curettings, and it is unclear if

they actually contribute to abnormal uterine bleed-ing. They are thought to be the ‘graveyards’ of priorimplantation sites, often remote. The nodules andplaques are of low cellularity, predominantly com-posed of eosinophilic material (Fig. 11). A potentialpitfall, although rare, are atypical placental site nod-ules (APSN) (22). On one series, 3/21 lesions termedAPSN were associated with GTD on follow-up,developing into either PSTT or ETT. These patientsdid not have elevated beta-hCG. The histologic fea-tures of the APSN were intermediate between PSNand PSTT/ETT, with lesions larger than the usualPSN, which rarely exceed 4 mm. APSNs also hadincreased cellularity, mild atypia, and an increasedki-67 index compared to usual PSN (22).

Exaggerated placental site

Exaggerated implantation sites occur at the time ofa gestation, either normal or molar. It is comprisedof implantation site trophoblasts infiltratingdecidua and between bundles of uterine smoothmuscle. When seen in smooth muscle, particularlyon curettings and if the chorionic villi have beenpassed, the concern of PSTT is sometimes raised.Exaggerated placental site (EPS) (Fig. 12) has moremultinucleation than PSTT, and a ki-67 index of 0,

Table 2. Immunohistochemistry of lesions of extravillous trophoblast

Cell type CK18AE1/3EMA

Mel-CAM Inhibin hPL PLAP hCG P63 Ki-67

EPS Implantation EVT + + + + Rare � 0PSN Chorionic EVT + occ +/� occ + occ + 5%PSTT Implantation EVT + + + + +/� Rare � 14%ETT Chorionic EVT + occ +/� occ occ occ +++ 18%

EPS, exaggerated placental site; PSN, placental site nodule; PSTT, placental site trophoblastic tumor; ETT, epithelioid tro-phoblastic tumor; EVT, extravillus trophoblast; occ, occasional.

Fig. 11. Placental site nodule-This nodule found in acurettage demonstrates the low cellularity of these lesions.


HELLER

as opposed to PSTT. EPS does not form masseslike PSTT. In addition, the presence of chorionicvilli helps establish the diagnosis.

Placental site trophoblastic tumor

Placental site trophoblastic tumor is composed ofextravillous trophoblast of implantational type. It ismonophasic (Fig. 13), unlike the biphasic chorio-carcinoma, and dissects between bundles of smoothmuscle, where it can resemble EPS. In addition tothe presence of a ki-67 index of about 14%, PSTTis more extensive than EPS, and PSTT is massforming. PSTT is usually non-aggressive in behav-ior, but metastatic disease has occurred. PSTT ismore chemoresistant than choriocarcinoma.

Epithelioid trophoblastic tumor

Similar in behavior to PSTT, ETT is derived fromchorionic type extravillous trophoblast, and so hasa different immunoprofile than PSTT (Table 2). Animportant characteristic is the presence of geo-graphic necrosis (Fig. 14). Like PSTT, it tends tobe chemoresistant.

Extrauterine lesions of extravillous trophoblasts:

Lesions of extravillous trophoblast can rarely occurin the fallopian tube, ovary or in paratubal tissue.This must be kept in mind for these rare lesions,and they should be distinguished from non-gesta-tional choriocarcinoma in the adnexal region. Aclassification system for these rare lesions has beendevised (23).

Future directions in GTD

Although still experimental, some feasibility hasbeen established for identifying GTD and distin-guishing it from non-gestational hCG-secretingtumors by evaluating circulating cell-free DNA(24).

CONCLUSIONS

Gestational trophoblastic disease continues to be achallenge for pathologists. Familiarity with the

Fig. 12. Exaggerated placental site. EVT is seen dissectingbetween smooth muscle bundles. Many of the cells aremultinucleated (inset).

Fig. 13. Placental site trophoblastic tumor-The lesion iscomprised of a monomorphous population of extravilloustrophoblast.

Fig. 14. Epithelioid trophoblastic tumor. The lesionshows geographic necrosis, and is comprised of amonomorphic proliferation of EVT (lower right), with aki-67 index (upper right), in the range of 18%.



histology and use of adjunctive techniques for unu-sual variants and early lesions will assist in the eval-uation of these specimens.

DISCLOSURES

None.

REFERENCES

1. Ngan HY, Seckl MJ, Berkowitz RS, Xiang Y, GolfierF, Sekharan PK, et al. Update on the diagnosis andmanagement of gestational trophoblastic disease. Int JGynaecol Obstet 2015;131(Suppl 2):S123–6.

2. Lurain JR. Gestational trophoblastic disease I: epi-demiology, pathology, clinical presentation and diag-nosis of gestational trophoblastic disease, andmanagement of hydatidiform mole. Am J ObstetGynecol 2010;203:531–9.

3. Lurain JR. Gestational trophoblastic disease II: classi-fication and management of gestational trophoblasticneoplasia. Am J Obstet Gynecol 2011;204:11–8.

4. Sebire NJ, Savage PM, Seckl MJ, Fisher RA.Histopathological features of biparental completehydatidiform moles in women with NLRP7 muta-tions. Placenta 2013;34:50–6.

5. Keep D, Zaragoza MV, Hassold T, Redline RW.Very early complete hydatidiform mole. Hum Pathol1996;27:708–13.

6. Sebire NJ. Histopathological diagnosis of hydatidi-form mole: contemporary features and clinical impli-cations. Fetal Pediatr Pathol 2010;29:1–16.

7. Buza N, Hui P. Immunohistochemistry and otherancillary techniques in the diagnosis of gestationaltrophoblastic diseases. Semin Diagn Pathol 2014;31:223–32.

8. Castrillon DH, Sun D, Weremowicz S, Fisher RA,Crum CP, Genest DR. Discrimination of completehydatidiform mole from its mimics by immunohisto-chemistry of the paternally imprinted gene productp57KIP2. Am J Surg Pathol 2001;25:1225–30.

9. Ronnett BM, DeScipio C, Murphy KM. Hydatidi-form moles: ancillary techniques to refine diagnosis.Int J Gynecol Pathol 2011;30:101–16.

10. Pawoo N, Heller DS. Placental mesenchymal dys-plasia. Arch Pathol Lab Med 2014;138:1247–9.

11. Wilson Y, Bharat C, Crook ML, Kee AR, Peverall J,Ruba S, et al. Histological comparison of partialhydatidiform mole and trisomy gestation specimens.Pathology 2016;48:550–4.

12. Khong TY. Chorangioma with trophoblastic prolifer-ation. Virchows Arch 2000;436:167–71.

13. Jiao L, Ghorani E, Sebire NJ, Seckl MJ. Intraplacen-tal choriocarcinoma: systematic review and manage-ment guidance. Gynecol Oncol 2016;141:624–31.

14. Ballieux BE, Weijl NI, Gelderblom H, van Pelt J,Osanto S. False-positive serum human chorionic gona-dotropin (HCG) in a male patient with a malignantgerm cell tumor of the testis: a case report and review ofthe literature. Oncologist 2008;13:1149–54.

15. Gonzalez AB, Syrios P, Gadisseur R, et al. Persistentlow levels of serum hCG due to heterophilic mouseantibodies: an unrecognized pitfall in the diagnosis oftrophoblastic disease. Gynecol Endocrinol 2016;32:439–41.

16. Levsky ME, Handler JA, Suarez RD, Esrig ET.False-positive urine beta-HCG in a woman with atubo-ovarian abscess. J Emerg Med 2001;21:407–9.

17. Anderson Z, Larson E, Khan M, Bell M. False nega-tive urine pregnancy testing with complete molar preg-nancy: an example of the hook effect. S D Med2016;69:55–7.

18. Cole LA, Muller CY. Hyperglycosylated hCG in themanagement of quiescent and chemorefractory gesta-tional trophoblastic diseases. Gynecol Oncol 2010;116:3–9.

19. Qian XG, Chen LL, Li BH, Cheng XD, Wan XY.Long term outcome of patients with persistent low-level elevation of human chorionic gonadotropin. JObstet Gynaecol Res 2016;42:694–700.

20. Lu B, Zhang X, Liang Y. Clinicopathologic analysis ofpostchemotherapy gestational trophoblastic neoplasia:an entity overlapping with epithelioid trophoblastictumor. Int J Gynecol Pathol 2016;35:516–24.

21. Cierna Z, Varga I, Danihel L Jr, Kuracinova K, Jane-gova A, Danihel L. Intermediate trophoblast – a dis-tinctive, unique and often unrecognized population oftrophoblastic cells. Ann Anat 2016;204:45–50.

22. Kaur B, Short D, Fisher R, Savage PM, Seckl MJ,Sebire NJ. Atypical placental site nodule (APSN) andassociation with malignant gestational trophoblasticdisease, a clinicopathologic study of 21 cases. Int JGynecol Pathol 2015;34:152–8.

23. Baergen RN, Rutgers J, Young RH. Extrauterinelesions of intermediate trophoblast. Int J GynecolPathol 2003;22:362–7.

24. Openshaw MR, Harvey RA, Sebire NJ, et al. Circu-lating cell free DNA in the diagnosis of trophoblastictumors. E Biomed 2015;4:46–52.

25. Shih I-M, Mazur MT, Kurman RJ. Gestational tro-phoblastic tumors and related tumor-like lesions. In:Kurman RJ, Ellenson LH, Ronnett BM, editors.Blaustein’s Pathology of the Female Genital Tract,6th ed., New York: Springer, 2011: 1085–8.


HELLER

Update on the pathology of gestational trophoblastic disease€¦ · Gestational trophoblastic...

Documents

Transcript of Update on the pathology of gestational trophoblastic disease€¦ · Gestational trophoblastic...