Intraplacental Gestational Neoplasms: A Review of Clinically Relevant Diagnostically Challenging Lesions

Choriocarcinoma is an uncommon, aggressive malignant trophoblastic neoplasm that can be gestational or nongestational in origin. Nongestational choriocarcinoma includes both germ cell origin and the rare association with somatic carcinomas. Distinction between gestational and nongestational choriocarcinoma is essential for treatment and prognosis according to international standards developed by the International Federation of Gynecology and Obstetrics (FIGO) and the World Health Organization.⁶  Gestational choriocarcinoma (GC) may arise from any gestational event, with wide geographic variation in incidence and precursors.⁷ 

IPC is a rare type of gestational trophoblastic neoplasia in which choriocarcinoma is detected within the placenta at any stage of gestation. First described in 1949 using the term “chorionepitheloma” by Douglas and Otto,⁸  then in 1963 by Driscoll⁹  as an “incidental choriocarcinoma” in a placenta from a term delivery, there are now more than 40 publications in the English language literature reporting IPC.^9–59  Additional reports have been published in French, German, and Japanese. Case reports and brief communications with or without literature review, letters to the editor, and inclusion of IPCs in ancillary and molecular evaluation of gestational trophoblastic neoplasia cohorts make up the body of IPC literature to date. Awareness of IPC may be increasing as reflected by an uptick in more recent publications (Figure 1), some of which presume IPC when presentation is temporally proximate to delivery but the placenta was discarded.^9,60–66  Still, fewer than 100 patients have been reported worldwide.

IPC remains relatively unknown in clinical and pathology domains. Recent clinical review articles continue to describe gestational choriocarcinoma as an entity defined by the absence of chorionic villi,^6,62,66  effectively denying the occurrence of IPC. This diagnostic feature appears to rely on an outdated understanding that GC originates in a residual adherent normal trophoblast following pregnancy that then has undergone or will undergo malignant transformation. This errant developmental theory has been repeatedly argued against, first by Brewer and Mazur¹²  in 1981, followed by Fox and Laurini¹⁵  in 1988, Hallam et al¹⁶  in 1990, and Duleba et al¹⁸  in 1992, all noting that the dogma requiring “absence of villi” to diagnose choriocarcinoma interferes with correct IPC diagnosis. Case studies at a center for gestational trophoblastic disease detected IPC in 0.04% of gestational trophoblastic neoplasia.²⁸  More recently, choriocarcinoma existing with or related to an otherwise normal term pregnancy was estimated at 1 in 50 000 to 100 000 term pregnancies^54,67  and 1 in 160 000 term pregnancies.^25,64  With nearly 4 million known births per year in the United States alone, this translates to at least 25 and up to 80 newly detected cases of IPC annually. Several reports postulate that this entity is likely to be underdiagnosed.^12,28,46 

Affected patients spanned childbearing age range, from 16 to 47 years, with a mean of 29 years and a median of 28.75 years. Gestational age at presentation ranged from 16 to 40 weeks.^9–59  Most reported patients were multiparous (G2, 26.1%; G3, 13.0%; G4 and greater, 21.3%; gravida not reported in 8%).^* Previous history of 1 or more abortions (spontaneous or other) was noted in in nearly 65% of the multiparous patients.^† Only 2 patients had previous history of molar gestation.^13,44  IPC occurring in primigravids, during this first known pregnancy accounts for the remaining 31.6% of cases.^‡

Nearly half of patients with histologically confirmed IPC presented with signs and symptoms related to metastasis: vaginal bleeding, anemia, melena, dyspnea, hemoptysis, headache, seizures, paresthesia. Metastatic disease was present at diagnosis in 42% of patients reported, and it often prompted extensive placental examination.^§ The remaining patients were asymptomatic or presented with abnormal features of pregnancy in the mother (eg, anemia, gestational hypertension, preterm labor), the fetus (eg, intrauterine growth restriction, reduced fetal movement, fetal hydrops, intrauterine fetal demise), or both (eg, massive maternal fetal hemorrhage, maternal fetal transfusion, disseminated intravascular coagulopathy). In some cases IPC was initially not identified, and it was only detected when the placenta was reexamined following persistent postpartum vaginal bleeding or endometrial evacuation.^12,15,40,51 

Clinical and laboratory evaluation following delivery in reported IPC cases was not performed according to standardized protocol but was guided by individual presenting features for each case. The most consistently reported abnormality was persistently elevated postpartum serum or urine β-human chorionic gonadotropin (B-HCG), seen in nearly all patients. Most patients with a diagnosis of IPC or who presented with symptoms of pulmonary, abdominal, or central nervous system metastasis were staged with radiologic evaluation that included chest radiography, abdominal and chest computed tomography scan, and magnetic resonance imaging to determine extent of metastatic disease.

When infants are delivered with anemia and pallor, raising suspicion for fetal-maternal hemorrhage/transfusion, the most consistently reported laboratory evaluations were hemoglobin, hematocrit, and complete blood counts of both mother and infant. The Kleihauer-Betke test, flow cytometry, and maternal hemoglobin electrophoresis (to detect fetal red blood cells or hemoglobin F) were variably performed. Evaluation for disseminated intravascular coagulopathy was reported for selected cases.

Examination of uterine curettage and surgical excision of more widely disseminated lesions confirmed metastases in several cases.^{8,11,12,31,33,34,43}  Early reports in which the disease course was uniformly fatal were evaluated in most instances with autopsy.^{8,10–12,14,19}  Fetopsy is often performed in cases of IPC-related intrauterine fetal demise (IUFD) or perinatal death,^38,44  and, rarely, autopsy more recently has confirmed undetected or untreated disease.⁵³ 

Affected placentas were referred for pathology examination for reasons other than suspicion of IPC, unless metastases were known at delivery. In 90% of cases, the fetal and maternal surfaces were grossly unremarkable. The rare maternal surface abnormalities included a localized yellow-tan waxy appearance similar to fibrin deposition,²¹  and areas of hemorrhage and compressing clot.¹⁴  A single case in which there was both chorangioma and IPC described a 4-cm lesion bulging from the fetal surface (chorangioma), distinct from a parenchymal 2.5-cm yellow and hemorrhagic lesion (IPC) that was able to be seen only on cut surface examination.²³ 

In more than 40% cases, the cut surface was grossly normal, with the area of IPC discovered incidentally at microscopic examination.^∥ IPC lesions may be solitary or multiple. Lesion size was reported in 83.6% of cases, and it ranged from microscopically detectable (9.0% of lesions) only to greater than 5 cm (6.0%). The remaining single or multiple grossly visible lesions were distributed as follows: less than 1 cm (19.4%), 1 to 2 cm (20.9%), 2 to 2.9 cm (19.4%), and 3 to 4.9 cm (9.0%). Size does not predict metastatic behavior. Most patients had disease limited to the placenta (58%), but metastasis was common (42%) and occurred with all sizes of tumor, including those only microscopically detected.^9–59 

When visible, the most common appearance resembled an infarct.^¶ Other lesions were described as hemorrhagic,^# including some with central necrosis,^{12,16,17,35,38,42,57}  reddish areas resembling fresh infarct,^14–16,29  numerous white-tan lesions focally abutting the maternal surface or forming infarctlike plaque,^20,21  yellow-tan,^{17,18,22–24,29}  sharply circumscribed,^22,24  or friable lesions.^16,17 

Across the series, an “atypical-appearing” infarct with heterogeneous areas of apparent necrosis or hemorrhage with a peripheral rim (Figure 2, A) is a repetitive theme which may guide detection and sampling for IPC. This gross appearance may correlate with the described histologic feature of “centrifugal growth”²⁰  in which the leading edge of the tumor is viable; however, the central area has collapsed with hemorrhage, infarction, or necrosis (Figure 2, B).

Cellular components of normal chorionic villi include cytotrophoblasts (CTs, which are generally indistinct during the third trimester), intermediate trophoblast (IT), syncytiotrophoblast (SCT), capillary endothelial cells, stromal mesenchymal cells, and Hoffbauer cells (histiocytes).⁷ 

IPC histologically varies in part related to the size of the detected lesion, but it is nonetheless defined by the characteristic biphasic or triphasic appearance of choriocarcinoma. Within smaller lesions, the affected villi display several layers of loosely arranged, atypical CTs and/or ITs with enlarged, hyperchromatic, and sometimes bizarre nuclear features, with atypical surface SCT (Figure 2, C). The combination of features has been described as resembling the implantation cap seen in early gestations^1,7  but with marked cytologic atypia; readily evident, sometimes abnormal mitoses; and apoptosis within the proliferating trophoblast. A junction between normal-appearing, nonproliferative villi and IPCs as well as atypical and normal trophoblasts in the same villus has been described.^{11–15,17,22,31}  Larger lesions display destruction of villi by tumor masses in clusters or sheets of biphasic or triphasic atypical cells with marked nuclear pleomorphism and readily identifiable mitoses, including atypical forms (Figure 2, D). Blood lakes and pseudovascular channels are haphazardly dispersed within the choriocarcinoma, and at higher magnification SCT lining cells are apparent. There is a striking absence of neovascularization in association with the mass, related to the tumor’s use of vasculogenic mimicry to establish vascular supply.⁶⁸  Vasculogenic mimicry is a general phenomenon of choriocarcinoma, recapitulating to some degree the vascular invasion and endothelial mimicry that occurs following implantation.⁶⁸  Extensive intratumoral hemorrhage, infarction, and necrosis are common.^{12,16,17,35,38,42,57}  Similar to smaller lesions, adjacent villi may display surface markedly atypical trophoblasts or be uninvolved and appear normal.^12,14,31 

Complete histologic evaluation involves careful examination and reporting of invasion into the involved or adjacent chorionic villus stroma and the chorionic villus capillaries, as well as into maternal vessels within the basal plate.

Although a biphasic or triphasic malignancy within the placenta may seem pathognomonic for choriocarcinoma, other high-grade malignancies that are capable of vasculogenic mimicry and metastasis can be considered.^67,69  Fetal malignancies with documented metastasis to the placenta include hepatoblastoma,^67,70  neuroblastoma, congenital acute leukemia, rhabdomyosarcoma, and malignant rhabdoid tumor.⁶⁷  Fetal metastases, both solid and hematologic malignancies, often display enlarged, edematous villi with neoplastic cells dilating the villous capillary lumen, rather than solid sheets of atypical cells.⁶⁷  Giant congenital melanocytic nevi with associated neurocutaneous melanocytosis and congenital melanoma with placental chorionic villi have been reported and are important to include within the differential diagnosis.⁷¹ 

Maternal malignancies metastatic to the placenta are rare but also can be considered in the IPC differential diagnosis.^67,72  Malignant melanoma is the most common metastasis to the placenta, followed by carcinoma of the breast or lung, with rare case reports of gastric and pancreatic adenocarcinoma.^67,72  Maternal sarcomas (angiosarcoma, epithelioid sarcoma, Ewing sarcoma, rhabdomyosarcoma) metastasizing to the placenta are noted in isolated reports.^67,73  Maternal metastases tend to occupy the maternal intervillous space adjacent to intact villi. Infarction or thrombus in IPC may give the appearance of commingled villi and metastatic malignancy.^67,72  The frequency of placental metastasis may be underestimated because the placenta may appear normal at gross examination and even with known maternal malignancy may not be sufficiently sampled for histologic examination.⁶⁷ 

Multifetal (twin) gestations with viable fetus and coexisting IPC are included within this review.^34,46,58  Histologic distinction of IPC in multifetal gestation from molar tissue present in partial⁷⁴  or complete hydatidiform mole with coexisting twin and higher order multifetal gestations^75–79  can be quite challenging. The distinction is especially difficult in first- and second-trimester gestations, and when placentas are fused or disrupted.⁷⁸  Evaluation with p57,⁸⁰  genotyping,⁸⁰  and consultation with colleagues with expertise in gestational trophoblastic neoplasia and placental pathology can clarify the diagnosis.

Immunohistochemistry is useful in establishing the diagnosis of choriocarcinoma, distinguishing the lesion from other diagnostic possibilities (ie, metastases) and also shedding light on disease development. Immunohistochemical markers that have been helpful in confirming trophoblastic differentiation include B-HCG,^7,15–17,20  human placental lactogen,⁷  GATA-3,⁸¹  and inhibin.⁸²  SALL4 has been demonstrated in choriocarcinoma of both gestational and nongestational origin.^83,84  Case reports in which immunohistochemistry had similar, reliable results across various laboratories included assays for cytokeratin (AE1/AE3, and others), B-HCG, and Ki-67/MIB-1. Cytokeratins are described as being variably to strongly positive in neoplastic SCT, as well as pCT (Figure 2, E) in 8 of 8 cases. B-HCG was reported to be strongly positive in both neoplastic and normal SCT (Figure 2, F) and variably positive within putative neoplastic CT (Figure 2, F) in 17 of 17 cases. Another hallmark of diagnosis is the reported high proliferative index (range, 75%–100%) in neoplastic mononuclear cells in 4 of 5 cases,^22,56,57,59  with a lower 15% index in only 1 of 5 cases.⁴⁸ 

There was significant variation in the findings across case reports for human placental lactogen, in which roughly half of cases showed positive staining in neoplastic cells, whereas half were negative.^** Other assays were performed only sporadically (CD34, collagen IV, SP-1, inhibin), which limits the usefulness of reporting the results of these stains.

To classify the various cell types within 36 gestational choriocarcinomas and various other gestational trophoblastic lesions, Mao et al⁸⁵  in 2007 applied Mel-CAM (CD146), HLA-G, MUC-4, and β-catenin. The findings in this in-depth study may be useful in distinguishing IPC from other gestational neoplastic diseases and in further understanding how choriocarcinoma develops; however, these antibodies are not widely used in routine practice.

Molecular evaluation of IPC is generally performed (1) for confirmation of gestational-type choriocarcinoma, (2) identification of type of gestation/index pregnancy for treatment stratification and prognostic information, and (3) research to elucidate the route to tumor development and metastasis.⁸⁶ 

Gestational choriocarcinomas are a genetically heterogeneous group of tumors. All gestational choriocarcinomas are distinct from the maternal genome and carry a paternal chromosome complement that is a unique product of the gestation.⁴⁷  Short tandem repeat (STR) analysis has been demonstrated to classify genomic origin (androgenetic versus biparental) and been used to confirm or refute the present gestation as the causative gestation in IPCs and gestational choriocarcinomas generally.⁸⁷ 

Unlike unique DNA sequences in the human genome, which exhibit a very low mutation rate, mutation rates in STR sequences are several orders of magnitude higher.^87,88  Analysis of several STR loci is used to determine if there is homology at specific loci, which indicates genotype matching. STR analysis performed on 3 IPCs demonstrated biparental genome, 2XX and 1 XY, genetically related to the normal placental villous tissue.⁴⁷  Genotyping analysis can be performed on formalin-fixed, paraffin-embedded tissue, and it is more accurate in determining the genomic origin and calculating time since index pregnancy than traditional clinical parameters are.

The route to oncogenesis for GC is unknown but thought to originate from trophoblast. Normal transient early trophoblast cells display a phenotype that supports rapid growth, rapid invasion, and vascular mimicry, resembling malignant cells.⁵⁰  Bulk placenta whole genome sequencing of 86 normal and abnormal nonneoplastic placentas found a genomic landscape akin to that of childhood cancer in terms of mutation burden and mutational imprints.⁴  Placentas can exhibit chromosomal aberrations that are absent from the fetus and be nonuniformly distributed within the placenta.⁴ 

IPC, albeit rare, represents an opportunity to evaluate the genetics and epigenetic alterations of choriocarcinoma in an early, de novo, untreated state. Fluorescent microsatellite genotyping (STR analysis), whole genome sequencing, methylation array hybridization, and bisulphite methylation analyses were performed on 1 IPC.⁵⁰  The genotype analysis confirmed that the tumor and healthy placenta shared the same genotype. Whole genome sequencing did not demonstrate any mutations of the common cancer-associated oncogenes or any other driver mutations. The tumor mutational load within specific somatic oncogenes was low. In contrast, the methylation pattern was markedly different between the tumor and the adjacent normal placenta. IPC displayed a genome-wide methylation profile resembling first-trimester chorionic villi.⁴⁷  The theory has been postulated that GC results from a block in development of early trophoblast cells, where they fail to lose the essentially malignant phenotype of the primitive trophoblast, possibly because of aberrant events in the normal progression of methylation.⁵⁰ 

Detection of IPC can provide early evaluation, risk stratification, and treatment for both mother and infant that could prove lifesaving. If weeks or months pass following delivery prior to GC detection, tumor burden can increase and opportunity for metastasis can occur. Detection at a later time interval confers a higher risk score, and potentially a higher stage and more significant morbidity and mortality from the cytotoxic chemotherapy used when higher stage disease is present. The cure rate for gestational choriocarcinoma is high; however, deaths related to the effects of chemotherapy for high-burden disease continue to be reported.⁴⁸ 

The true incidence of IPC is unknown. Placental examination was not performed in a significant subset of patients who presented with near-immediate postpartum GC, nor in infants in whom neonatal or congenital choriocarcinoma is detected.^{10,60–66,89} 

The current referral of placenta for pathology evaluation practice guidelines⁹⁰  and placenta gross examination practices are unlikely to detect all IPC.⁹¹  Abnormalities of the fetal or maternal surface that would be visible to the delivering clinicians, prompting referral, are uncommon in IPC. Many reported cases of IPC were “incidental,” detected during placental examination performed for other reasons without suspicion for GTN. Once referred to pathology, widely used gross examination methods may contribute to underdetection of lesions. Routine examination is to perform 1- to 2-cm–thick slices and then to examine the slices’ cut surface on each side of the slice through the placenta to evaluate for parenchymal lesions.^90,92  This approach has a high likelihood of missing lesions that are less than 1 cm in size,⁹¹  which accounts for approximately 20% of lesions. Detected lesions have a gross appearance resembling infarcts or areas of hemorrhage and thrombus. Current practice is to sample representative lesions, and this does not require sampling of all such lesions, which may further compromise rates of IPC detection.

Heightened suspicion for IPC when associated clinical features are present warrants more liberal sampling of grossly unremarkable parenchyma and any space-occupying lesions. Associated clinical features and the percentage occurring include any symptoms of metastatic disease (41.3%), peripartum vaginal bleeding (16.2%), IUFD (13.5%), fetal/maternal hemorrhage (12.2%), reduced fetal movement (10.8%), fetal distress (9.5%), preterm labor (5.4%), prior history of molar gestation or any GTD (2.7%), intrauterine growth restriction (IUGR) (2.7%), or concurrent partial molar gestation with viable twin gestation (1.3%). At the time of this report, only 1 patient has been reported with 2 sequential IPCs.⁴⁴  The first IPC case followed IUFD at 35 weeks with an IUGR fetus and trisomy 9. Thirteen months later, the placenta for a subsequent pregnancy was referred and examined very carefully. Neither placenta contained macroscopically visible lesions. Liberal sampling from each gestation resulted in histologic diagnosis. Genetic analysis to confirm causative choriocarcinoma on the second gestation was not performed.

The classic caveat that “the diagnosis of choriocarcinoma is negated by the presence of villous structures” remains within current literature^6,62  despite the fact that IPC has been demonstrated arising in term and preterm placentas among chorionic villi, and with partial involvement of intact villi. Neoplastic features in the trophoblast (proliferation, atypia, mitotic activity, formation of pseudovascular spaces with maternal or fetal red cells) surrounding villi within the placenta warrant consideration of the diagnosis of IPC, additional confirmation of cell types by immunohistochemistry, and referral for consultation when indicated. Routine genotyping of gestational choriocarcinomas at time of presentation may provide more refined management for affected patients.⁴⁷ 

The term “chorangiocarcinoma” was originally used by Jauniaux et al⁹³  in 1988. The proposed terminology defines a lesion containing atypical trophoblastic proliferation equivalent to that seen in IPC combined with marked vascular proliferation similar or equivalent to chorangioma. There are 7 English-language reports of “chorangiocarcinoma” and an eighth in German.^93–100  Two additional reports^101,102  use the terms “chorangioma with trophoblastic proliferation” or “chorangioma with atypical trophoblast proliferation” to report cases with histologic images and descriptions similar to those in the original publication by Jauniaux et al.⁹³  These reports were recently the subject of a systematic literature review.¹⁰³  One additional case exists within the author’s consultation files that is grossly similar and histologically equivalent to a subset of the reported cases (Figure 3, A and B).^97–99 

Following the first few “chorangiocarcinoma” reports, Khong in 2000¹⁰¹  performed single-institution retrospective examination to evaluate trophoblast proliferation and cellular features of chorangiomas to establish a “baseline”definition. Of 23 tumors evaluated, 65% met criteria proposed in previous reports advancing the term “chorangiocarcinoma.” Zero patients in Khong’s retrospective review had subsequent recurrence or metastases; thus, the term and disease significance remained uncertain. Later chorangiocarcinoma reports^98,99  included 2 patients in whom postpartum B-HCG failed to decrease to zero, and metastatic disease was detected within 90 days. Both were treated with chemotherapy and achieved completed remission with 1-year follow-up.^98,99  Intrauterine fetal demise, seen in 13.5% of IPC, was not observed in the 9 cases of chorangiocarcinoma. Eight of 9 reported infants delivered were alive and well with no evidence of disease. One infant delivered at 30 weeks prompted by maternal pre-eclampsia and multiple hypoechoic placental lesions died at 9 days of life with Type 1 respiratory distress syndrome and sepsis.¹⁰² 

Patients ranged from age 23 to 36, 33% primigravid, 55% multiparous, and 1 patient not specified.^93–100  There was a prior history of abortion in 33%, and no reported prior partial or complete molar gestations. One lesion was reported in a twin gestation,⁹⁴  the remaining in single intrauterine pregnancies.^93,95–100  In contrast to IPC, one-third of patients with chorangiocarcinoma had prenatal ultrasound–placental lesions, all suspected to be chorangiomas.^97,100,102  Despite the size of the suspected chorangiomas, no fetal anemia, cardiomyopathy, or presenting symptoms of fetal-maternal hemorrhage were reported. Clinical presentation varied, with 1 patient displaying each of the following: vaginal bleeding,⁹³  intrauterine growth retardation,⁹⁵  severe condyloma,⁹⁶  abnormal screening for alpha fetoprotein⁹⁶  or other markers used to detect trisomy 21,⁹⁷  macrosomia,⁹⁸  vaginal discharge and preterm premature rupture of membranes,¹⁰⁰  and pre-eclampsia.¹⁰²  Except for the single patient with vaginal bleeding, no patients in this group presented with clinical signs or symptoms suggesting metastasis.

Placentas were referred for examination because of visible abnormalities (6 of 9, including 3 of 9 with prior ultrasound detection of a lesion) and maternal or infant clinical indications (3 of 9). The fetal surface of 4 placentas was visibly abnormal. A fifth placenta was markedly enlarged (1200 g) with “grapelike clusters” visible at delivery.¹⁰²  Involved fetal surfaces contained well-demarcated nodules or protrusions just under^93,100  or bulging from^96–99  the chorionic plate. Fetal surface lesions ranged from 2.5 to 5.5 cm in single greatest dimension. Three lesions were described within 3 cm of the base of the umbilical cord,^96–98  and 2 as peripheral.^94,98  Single lesions predominated (8 of 9 cases), with only 1 patient having multiple lesions.¹⁰² 

On cut surface, the tumors ranged from 1.5 to 8.0 cm in single greatest dimension. Tumors were estimated to occupy as little as 2% to as much as 26% of the placenta. The lesions were firm in 8 of 9 cases, with variable color: white,⁹³  tan,^94,96  yellow-gray,⁹⁵  reddish brown with white internal nodules,⁹⁷  and grayish to yellow-white.^98,99  A single case was described as having a friable cut surface.¹⁰⁰ 

Early reports described a classic angiomatous pattern with a well-differentiated capillary pattern of benign endothelium supported by a dense network of stromal cells and fibrous tissue, as is generally seen in chorangiomas. Normal stem vessels and chorangiomas have an overlying layer of trophoblast in which generally only SCT is visible. In the described lesion, the outer layer of the tumor was formed by multiple layers of SCT admixed with CT (possibly intermediate trophoblast) that displayed marked nuclear pleomorphism similar to that seen in true trophoblastic neoplasms.⁹³  Subsequently, 4 reports displayed similar features,^94,95,99,102  noting circumferential abnormal, malignant-appearing cells, the presence of numerous mitoses, and striking vascular proliferation within the adjacent stroma. No stromal invasion or invasion into maternal vessels was evident. None of the patients had recurrence or metastasis after delivery of the placenta, rendering the significance of these findings uncertain.

Later reports of chorangiocarcinoma included a morphologically distinct variant in which the tumor appears as large nodules and solid sheets of biphasic and triphasic malignant-appearing trophoblasts with central necrosis, dystrophic calcification, and granular debris (Figure 3, C).^96–98,100  Vascular proliferation adjacent to the nodules similar to that of chorangioma and the previously described lesions was also striking (Figure 3, D). The possibility was raised that the angiomatous vessels of chorangioma were dilated and expanded by proliferating, neoplastic trophoblasts.⁹⁶  The features are somewhat reminiscent of an inverted papilloma, in which the epithelial cells face luminally. No stromal invasion or maternal vascular invasion was evident in the group of patients with this second morphologic pattern. This may indicate lesions along a spectrum, a collision tumor, or 2 separate, distinct entities. A total of 2 of 5 patients with the nodular, “inverted” architecture presented with postpartum metastases, both to the lung.^98,99 

The differential diagnosis for these lesions depends on which histologic pattern for “chorangiocarcinoma” is evident. When proliferating trophoblasts are limited to the external surface of chorangioma, the lesion needs to be distinguished from benign chorangioma with trophoblast proliferation. When the lesion displays the broad nodular form, evaluation for metastatic lesions from other sites (squamous cell carcinoma, melanoma, breast carcinoma, and others) is prudent.

Ancillary studies were limited to immunohistochemistry evaluation. Evaluation was variable across the 9 reports, using various antibodies as reported in the original manuscripts,^93–102  with results summarized in the Table. B-HCG can be diffusely positive (Figure 3, E) or restricted to trophoblasts. CD34 (Figure 3, F) can confirm the extent of vascular proliferation.

No patient material was subjected to STR genotypic analysis to confirm patient origin or other molecular evaluations, nor was any included in any related molecular study for gestational trophoblastic neoplasms.

“Chorangiocarcinoma” and “chorangioma with atypical trophoblast proliferation” are rare entities that appeared to follow a benign course when first reported in the literature. Two recent cases showed the metastatic behavior possible with these lesions.^98,99  Clarification of specific diagnostic criteria that would reliably differentiate chorangiomas with trophoblast proliferation (a common finding) from those with sufficiently atypical features to warrant diagnosis of “atypical trophoblast proliferation” or even chorangiocarcinoma will require additional retrospective reviews similar to the work of Khong¹⁰¹  and prospective studies using cases that are now recognized as within the spectrum of these lesions. The possibility exists that the terms “chorangiocarcinoma” and “chorangioma with atypical trophoblast proliferation” refer to 2 distinct lesions, lesions along a spectrum, or a collision of choriocarcinoma into a preexisting chorangioma. To date, metastasis has only occurred with the morphologic entity that displays an “inverted” pattern with broad nodules of malignant trophoblast surrounding a central area of necrosis within what resembles a chorangioma.

Like IPC, the understanding of oncogenesis and the elucidation of mechanisms that portend metastasis are not known. The paucity of chorangiocarcinoma cases prevents a clear understanding of the nature of these lesions and extrapolating available data to estimate risks posed to both mother and infant in subsequently detected lesions.

The geographic incidence of GC detected within and following term pregnancy is variable. Clinical scoring systems (FIGO) that aim to predict biologic behavior, specifically to identify patients who will be resistant to single-agent chemotherapy, remain imperfect, with one-third of “low-risk patients” developing disease resistance.¹⁰⁴  There is increasing awareness that most GCs following term pregnancy have developed within the placenta (IPC) during gestation, particularly those with a short interval between delivery and presentation.⁸⁹  When detected via postpartum metastasis, it is postulated that the vast majority of the primary lesions go unsampled in their de novo state within the placenta. Thus, current GC clinical understanding and research is largely based on the behavior of recurrences and metastases, rather than the de novo lesions.

Intraplacental gestational neoplasms include IPC and the rarer, less well-defined “chorangiocarcinoma” (also known as chorangioma with atypical trophoblast proliferation). The true incidence of these neoplasms is likely underestimated because calculations are biased toward the subset of patients who present with metastases or in whom metastases are detected. Early reports used the term “choriocarcinoma in situ” interchangeably with “intraplacental choriocarcinoma.” More than 40% of reported IPCs and 22% of chorangiocarcinoma patients present with or are discovered to have metastasis at the time of delivery. For this reason the term “choriocarcinoma in situ,” which portrays a benign course, is not favored.

IPC and the less frequent “chorangiocarcinoma” range from an isolated microscopic focus within macroscopically normal placentas to large, macroscopically visible lesions that occupy up to 25% of the placenta. Lesion size alone has been an ineffective predictor of biologic behavior (Figure 4) in IPC. Microscopic IPC lesions have metastasized widely to both mother and infant, whereas large lesions can be clinically silent and resolve spontaneously after the placenta is evacuated from the uterus. Lesion size greater than 5 cm for chorangiocarcinoma was present in both cases with metastatic disease. There are too few cases in the literature to reliably predict the biologic behavior of choriangiocarcinomas less than 5 cm in size.

Continued dissemination to clinicians and pathologists alike regarding the clinical features associated with IPC and “chorangiocarcinoma” will reinforce the need for submission of placentas to pathology for gross and histologic examination. Features associated with IPC and/or chorangiocarcinoma (metastatic choriocarcinoma detected during pregnancy, vaginal bleeding, placental mass detected at ultrasound, fetal-maternal hemorrhage or transfusion, fetal anemia, premature delivery, IUFD, IUGR) are all clinical indications to refer the placenta.⁹⁰  Heightened suspicion within pathology for these lesions can result in careful macroscopic assessment with sectioning at less than 1-cm intervals.⁹²  Given the macroscopic appearance of these lesions, more liberal sampling of submitted placentas, including areas resembling infarctions, intervillous thrombi, and nodules with characteristic features of chorangiomas, may assist in higher rates of detection.

In addition to earliest intervention, increased sampling and diagnosis of IPC and “chorangiocarcinoma” represent an opportunity to elucidate disease mechanisms and tumor characteristics that can portend a high risk of fetal-maternal hemorrhage, poor fetal outcomes, and metastatic behavior. B-HCG levels vary widely during pregnancy, making it an unreliable assay for primary detection of these lesions. Identification of other biomarkers related to malignant transformation or vascular mimicry with less-invasive test methods would be ideal for early diagnosis and intervention. With fewer than 100 IPCs estimated to occur annually, referral from community facilities and multicenter collaboration will be required to rapidly advance our understanding.