Standardized Classification of Lung Adenocarcinoma Subtypes and Improvement of Grading Assessment Through Deep Learning.

The histopathologic distinction of lung adenocarcinoma (LADC) subtypes is subject to high interobserver variability, which can compromise the optimal assessment of patient prognosis. Therefore, this study developed convolutional neural networks capable of distinguishing LADC subtypes and predicting disease-specific survival, according to the recently established LADC tumor grades. Consensus LADC histopathologic images were obtained from 17 expert pulmonary pathologists and one pathologist in training. Two deep learning models (AI-1 and AI-2) were trained to predict eight different LADC classes. Furthermore, the trained models were tested on an independent cohort of 133 patients. The models achieved high precision, recall, and F1 scores exceeding 0.90 for most of the LADC classes. Clear stratification of the three LADC grades was reached in predicting the disease-specific survival by the two models, with both Kaplan-Meier curves showing significance (P = 0.0017 and 0.0003). Moreover, both trained models showed high stability in the segmentation of each pair of predicted grades with low variation in the hazard ratio across 200 bootstrapped samples. These findings indicate that the trained convolutional neural networks improve the diagnostic accuracy of the pathologist and refine LADC grade assessment. Thus, the trained models are promising tools that may assist in the routine evaluation of LADC subtypes and grades in clinical practice.

Malignant peritoneal mesothelioma: prognostic significance of clinical and pathologic parameters and validation of a nuclear-grading system in a multi-institutional series of 225 cases.

Malignant peritoneal mesothelioma historically carried a grim prognosis, but outcomes have improved substantially in recent decades. The prognostic significance of clinical, morphologic, and immunophenotypic features remains ill-defined. This multi-institutional cohort comprises 225 malignant peritoneal mesotheliomas, which were assessed for 21 clinical, morphologic, and immunohistochemical parameters. For epithelioid mesotheliomas, combining nuclear pleomorphism and mitotic index yielded a composite nuclear grade, using a previously standardized grading system. Correlation of clinical, morphologic, and immunohistochemical parameters with overall and disease-free survival was examined by univariate and multivariate analyses. On univariate analysis, longer overall survival was significantly associated with diagnosis after 2000 (P = 0.0001), age <60 years (P = 0.0001), ECOG performance status 0 or 1 (P = 0.01), absence of radiographic lymph-node metastasis (P = 0.04), cytoreduction surgery (P < 0.0001), hyperthermic intraperitoneal chemotherapy (P = 0.0001), peritoneal carcinomatosis index <27 (P = 0.01), absence of necrosis (P = 0.007), and epithelioid histotype (P < 0.0001). Among epithelioid malignant mesotheliomas only, longer overall survival was further associated with female sex (P = 0.03), tubulopapillary architecture (P = 0.005), low nuclear pleomorphism (P < 0.0001), low mitotic index (P = 0.0007), and low composite nuclear grade (P < 0.0001). On multivariate analyses, the low composite nuclear grade was independently associated with longer overall and disease-free survival (P < 0.0001). Our data further clarify the interactions of clinical and pathologic features in peritoneal mesothelioma prognosis and validate the prognostic significance of a standardized nuclear-grading system in epithelioid malignant mesothelioma of the peritoneum.

Challenges in Ki-67 assessments in pulmonary large-cell neuroendocrine carcinomas.

AIMS: To gather the best available evidence regarding Ki-67% values in large-cell neuroendocrine carcinoma (LCNEC) and determine whether certain cut-off values could serve as a prognostic feature in LCNEC. METHODS AND RESULTS: Aperio ScanScope AT Turbo, eSlide Manager and ImageScope software (Leica Biosystems) were used to measure Ki-67% in 77 resected LCNEC diagnosed by World Health Organisation (WHO) criteria. Cases were stratified into six classes by 10% Ki-67 increments. Using the Kaplan-Meier method, overall (OS) and disease-free survivals (DFS) were compared by AJCC stage, by six Ki-67% classes and with Ki-67% cut-points ≥20% and ≥40%. Tumours were from 0.9 to 11.5 cm and pathological stages 1-3. The system measured Ki-67% positivity using 4072-44 533 tumour nuclei per case (mean 16610 ± 8039). Ki-67% ranged from 1 to 64% (mean = 26%; median = 26%). Only 16 (21%) tumours had Ki-67% ≥40%. OS ranged from 1 to 298 months (median follow-up = 25 months). DFS ranged from 1 to 276 months (median follow-up = 9 months). OS and DFS differed across AJCC stage (overall log-rank P = 0.038 and P = 0.037). However, neither OS nor DFS significantly correlated with Ki-67% when six or two classes were used with either ≥20% Ki-67 or ≥40% Ki-67 as cut-point. A literature review identified 14 reports meeting our inclusion criteria with ≥10 LCNEC. Reported Ki-67% ranged from 2% to 100%. Problems contributing to variability in Ki-67% measurements are discussed. CONCLUSION: Our findings caution against a blanket use of 20%, 40% or other Ki-67% cut-points for LCNEC diagnosis or prognostication.

Membranous HEG1 expression is a useful marker in the differential diagnosis of epithelioid and biphasic malignant mesothelioma versus carcinomas.

Sialylated HEG1 has been reported as a highly specific and sensitive mesothelioma marker but a comprehensive evaluation of its expression in carcinomas in different organs, various sarcomas and reactive mesothelial proliferations has not been reported. The aim of this study was to evaluate the clinical applicability of HEG1 as a marker in the diagnosis of mesothelioma. HEG1 immunoreactivity was evaluated in whole sections of 122 mesotheliomas, 75 pulmonary carcinomas, 55 other carcinomas, 16 mesenchymal tumors, and 24 reactive mesothelial proliferations and in tissue microarrays containing 70 epithelioid (EM), 36 biphasic (BM), and 2 sarcomatoid mesotheliomas (SM). In whole sections and tissue microarrays, respectively, membranous HEG1 was expressed in 93.0% and 85.5% of EM, 81.3% and 69.4% of BM, 0% and 0% of SM. HEG1 was not expressed in pulmonary adenocarcinomas. HEG1 was expressed as cytoplasmic immunoreactivity in pulmonary squamous cell carcinomas (21.7%). Membranous HEG1 staining was seen in ovarian carcinomas (66.7%), thyroid carcinomas (100%), reactive conditions (16.7%), and mesenchymal tumors (18.8%). The sensitivity of membranous HEG1 expression to distinguish EM/BM from all carcinomas was 88.8%. The specificity for the differential diagnosis between EM/BM and all carcinomas and pulmonary carcinomas was 92.3% and 98.7%, respectively.

Clinico-pathologic findings in patients with median arcuate ligament syndrome (celiac artery compression syndrome).

Median Arcuate Ligament Syndrome (MALS) is a rare entity characterized by severe post-prandial epigastric pain, nausea, vomiting, and/or weight loss. Symptoms have been attributed to vascular compression (celiac artery compression syndrome, CACS), but it remains controversial whether they could be secondary to neural compression. Literature review identified rare description of pathologic findings in surgery journals. The clinico-pathologic findings of four MALS patients who underwent robotic or laparoscopic surgery in our hospital are described. All our patients were female with a median age of 32.5 (range 25-55 years), and a median BMI of 23.5 kg/m2. They presented with chronic often post-prandial abdominal pain (4/4), nausea (3/4), emesis (2/4), anorexia (1/4), and weight loss (1/4). Two patients had a history of Crohn's disease. At intraoperative exploration, the celiac artery and adjacent nerves and ganglia were encased and partially compressed by fibrotic tissue in each patient. In each case laparoscopic excision of fibrotic tissue, celiac plexus and ligament division and was performed; celiac plexus nerve block was also performed in one patient. After surgical intervention, symptoms improved in three of the patients whose specimens show periganglionic and perineural fibrosis with proliferation of small nerve fibers. Our findings support neurogenic compression as a contributing factor in the development of pain and other MALS symptoms, and favor the use of MALS rather than CACS as diagnostic terminology. To further study the pathogenesis of this unusual syndrome, surgeons should submit all tissues excised during MALS procedures for histopathologic examination.

Localized malignant mesothelioma, an unusual and poorly characterized neoplasm of serosal origin: best current evidence from the literature and the International Mesothelioma Panel.

Localized malignant mesotheliomas (LMM) is an uncommon and poorly recognized neoplasm. Its pathologic diagnosis is often surprising in patients with serosal/subserosal based localized tumors that are clinically suspicious for metastatic lesions or primary sarcomas. Once a tumor is diagnosed as "mesothelioma", LMM is often mistaken for diffuse malignant mesothelioma (DMM). Best currently available evidence about LMM was collected from the literature and cases diagnosed by members of the International Mesothelioma Panel (IMP). One hundred and one (101) LMM have been reported in the English literature. Patients had localized tumors with identical histopathologic features to DMM. Patients ranged in age from 6 to 82 years; 75% were men. Most (82%) of the tumors were intrathoracic. Others presented as intrahepatic, mesenteric, gastric, pancreatic, umbilical, splenic, and abdominal wall lesions. Tumors varied in size from 0.6 to 15 cm. Most patients underwent surgical resection and/or chemotherapy or radiation therapy. Median survival in a subset of patients was 29 months. Seventy two additional LMM from IMP institutions ranged in age from 28 to 95 years; 58.3% were men. Sixty tumors (83.3%) were intrathoracic, others presented in intraabdominal sites. Tumors varied in size from 1.2 to 19 cm. Median survival for 51 cases was 134 months. Best evidence was used to formulate guidelines for the diagnosis of LMM. It is important to distinguish LMM from DMM as their treatment and prognosis is different. A multidisciplinary approach is needed for the diagnosis of LMM as it shows identical histopathology and immunophenotype to DMM.

Prognosis in pathology: Are we "prognosticating" or only establishing correlations between independent variables and survival? A study with various analytics cautions about the overinterpretation of statistical results.

Survival data from 225 patients with resected pulmonary typical carcinoids were analyzed with Kaplan-Meier statistics (K-M) and "deep learning" methods to illustrate the difference between establishing "correlations" and "prognostications". Cases were stratified into G1 and G2 classes using a ≤5% Ki-67% cut-point. Overall survival, number of patients at risk and 95% confidence intervals (CI) were estimated for the two classes. Seven neural network models (NN) were developed with GMDH Shell 3.8.2 and Statgraphics Centurion 18.1 software, using variable prior probabilities and different numbers of training vs testing cases. The NNs used age, sex, and pTNM, G1 and G2 as input neurons and "alive" and "dead" as output neurons. Areas under the curve (AUC) and other performance measures were evaluated for all models. Log-rank test showed a significant difference in overall survival between G1 and G2 (p < 0.001). However, 95% CI estimates showed considerable variability in survival at different time intervals. Including the number of patients at risk at different time intervals showed that most G2 patients had been censored by 100 weeks. The NN models provided variable "prognostications", with AUC ranging from 0.5 to 1 and variability in the sensitivity, specificity, and other performance measures. The results illustrate the limitations of survival statistics and NNs in predicting the prognosis of individual patients. The need for pathologists not to overinterpret the finding of significant correlations as "prognostic" or "predictive" for individual patients is discussed.

Pathologists should probably forget about kappa. Percent agreement, diagnostic specificity and related metrics provide more clinically applicable measures of interobserver variability.

Kappa statistics have been widely used in the pathology literature to compare interobserver diagnostic variability (IOV) among different pathologists but there has been limited discussion about the clinical significance of kappa scores. Five representative and recent pathology papers were queried using clinically relevant specific questions to learn how IOV was evaluated and how the clinical applicability of results was interpreted. The papers supported our anecdotal impression that pathologists usually assess IOV using Cohen's or Fleiss' kappa statistics and interpret the results using some variation of the scale proposed by Landis and Koch. The papers did not cite or propose specific guidelines to comment on the clinical applicability of results. The solutions proposed to decrease IOV included the development of better diagnostic criteria and additional educational efforts, but the possibility that the entities themselves represented a continuum of morphologic findings rather than distinct diagnostic categories was not considered in any of the studies. A dataset from a previous study of IOV reported by Thunnissen et al. was recalculated to estimate percent agreement among 19 international lung pathologists for the diagnosis of 74 challenging lung neuroendocrine neoplasms. Kappa scores and diagnostic sensitivity, specificity, positive and negative predictive values were calculated using the majority consensus diagnosis for each case as the gold reference diagnosis for that case. Diagnostic specificity estimates among multiple pathologists were > 90%, although kappa scores were considerably more variable. We explain why kappa scores are of limited clinical applicability in pathology and propose the use of positive and negative percent agreement and diagnostic specificity against a gold reference diagnosis to evaluate IOV among two and multiple raters, respectively.

Pathology in the era of "Personalized Medicine": The need to learn how to integrate multivariate immunohistochemical and "omics" data with clinicopathologic information in a clinically relevant way".

"Personalized medicine" has been proposed as a new paradigm for patient care that, based on the integration of genomics and other "omics" data with clinical and other multidisciplinary information, promises early disease detection, improved outcomes and reduced side effects to therapies. Pathologists have become important participants in this new approach as the guardians of tissues and experts in the performance of molecular and other laboratory tests. Large amounts of new laboratory data in multiple neoplasms and other entities are being reported but there has been limited discussion about how best to evaluate the clinical significance of this information and how to integrate it into currently available diagnostic and therapeutic modalities. This article introduces a variety of epistemological problems presented by the "personalized medicine" paradigm and briefly discusses various topics that will be evaluated in further detail in future articles of this new series on Evidence-Based Pathology.

Carcinoid tumors of the thymus and Cushing's syndrome: Clinicopathologic features and current best evidence regarding the cell of origin of these unusual neoplasms.

It is uncertain whether thymic neuroendocrine tumors (NET) associated with Cushing's syndrome (CS) produce corticotropin-releasing hormone (CRH) and adrenocorticotropin hormone (ACTH) and whether the thymus contains ACTH and/or CRH cells that could originate NET. The clinicopathologic features of 5 typical (TC) and 6 atypical carcinoids (ATC), 10 additional non-neoplastic thymi, 6 adrenal glands with bilateral nodular hyperplasia and 8 adrenal cortical adenomas were reviewed. Representative slides were immunostained for ACTH and CRH. Four (36.4%) of the 11 patients had CS. The incidence of Masaoka stage IV was higher (p < 0.0001) in patients with ATC than TC. Only 2 (18.1%) of the 11 patients were alive at follow-up. Ten NET were CRH immunoreactive and 6 were ACTH immunoreactive. Thymic NET with CS exhibited stronger immunoreactivity for ACTH and CRH than those without CS. Non-neoplastic thymi exhibited scattered ACTH and CRH immunoreactive cells. Normal adrenal cortex and glands with bilateral nodular hyperplasia showed diffuse CRH immunoreactivity while adrenal adenomas showed no or only focal CRH immunoreactivity. Literature review showed no association between thymic NET and adrenal adenomas. The thymus contains CRH and ACTH immunoreactive cells that are probably the origin of thymic NET. Neoplasms associated with CS exhibit strong immunoreactivity for both hormones, suggesting that CRH probably plays a role in the pathogenesis of CS. As adrenals with bilateral nodular hyperplasia exhibit diffuse CRH immunoreactivity and adrenal cortical adenomas either lack this finding or show few immunoreactive cells, this marker may be useful to distinguish these lesions.

The use of Ki-67 labeling index to grade pulmonary well-differentiated neuroendocrine neoplasms: current best evidence.

Although Ki-67 labeling index (Ki-67%) is not a diagnostic or grading criterion in the World Health Organization classification of pulmonary carcinoid tumor, oncologists often request this test. A survey was administered at a North American Society for Neuroendocrine Tumors meeting to understand how Ki-67% is used in oncologic practices. A systematic literature review was performed to gather best evidence regarding the use of Ki-67%. Consecutive pulmonary carcinoids were stratified into pulmonary typical carcinoids with Ki-67% <5% (group A, n = 187), typical carcinoids with Ki-67% ≥5% (group B, n = 38) and atypical carcinoids irrespective of Ki-67% (group C, n = 31). Overall survival, progression-free survival, recurrence proportions and time to recurrence were compared, by group, using the log-rank test, chi-square statistics and ANOVA, respectively. Our survey confirmed that Ki-67% is frequently used by specialists caring for these patients. Ki-67% of 1-7% significantly correlated with overall survival in the literature but we found no information about Ki-67% cut-off values that would accurately distinguish pulmonary typical from atypical carcinoids or estimate the prognosis of patients stratified by World Health Organization diagnosis and Ki-67% cut-off. Overall survival was significantly different in our 3 patient groups (p < 0.001), with survival probabilities decreasing from groups A to C. Progression-free survival was significantly longer in group A than B (p < 0.007). Our results support the concept that by combining World Health Organization diagnosis and Ki-67%, pulmonary carcinoids can be stratified into 3 grades: G1 (typical carcinoids with Ki-67% <5), G2 (typical carcinoids with Ki-67% ≥5%) and G3 (atypical carcinoids) with different prognoses.

Nuclear grade and necrosis predict prognosis in malignant epithelioid pleural mesothelioma: a multi-institutional study.

A recently described nuclear grading system predicted survival in patients with epithelioid malignant pleural mesothelioma. The current study was undertaken to validate the grading system and to identify additional prognostic factors. We analyzed cases of epithelioid malignant pleural mesothelioma from 17 institutions across the globe from 1998 to 2014. Nuclear grade was computed combining nuclear atypia and mitotic count into a grade of I-III using the published system. Nuclear grade was assessed by one pathologist for three institutions, the remaining were scored independently. The presence or absence of necrosis and predominant growth pattern were also evaluated. Two additional scoring systems were evaluated, one combining nuclear grade and necrosis and the other mitotic count and necrosis. Median overall survival was the primary endpoint. A total of 776 cases were identified including 301 (39%) nuclear grade I tumors, 354 (45%) grade II tumors and 121 (16%) grade III tumors. The overall survival was 16 months, and correlated independently with age (P=0.006), sex (0.015), necrosis (0.030), mitotic count (0.001), nuclear atypia (0.009), nuclear grade (<0.0001), and mitosis and necrosis score (<0.0001). The addition of necrosis to nuclear grade further stratified overall survival, allowing classification of epithelioid malignant pleural mesothelioma into four distinct prognostic groups: nuclear grade I tumors without necrosis (29 months), nuclear grade I tumors with necrosis and grade II tumors without necrosis (16 months), nuclear grade II tumors with necrosis (10 months) and nuclear grade III tumors (8 months). The mitosis-necrosis score stratified patients by survival, but not as well as the combination of necrosis and nuclear grade. This study confirms that nuclear grade predicts survival in epithelioid malignant pleural mesothelioma, identifies necrosis as factor that further stratifies overall survival, and validates the grading system across multiple institutions and among both biopsy and resection specimens. An alternative scoring system, the mitosis-necrosis score is also proposed.

Dataset for reporting of thymic epithelial tumours: recommendations from the International Collaboration on Cancer Reporting (ICCR).

AIMS: The International Collaboration on Cancer Reporting (ICCR) is a not-for-profit organization formed by the Royal Colleges of Pathologists of Australasia and the United Kingdom, the College of American Pathologists, the Canadian Association of Pathologists-Association Canadienne des Pathologists in association with the Canadian Partnership Against Cancer, and the European Society of Pathology. Its goal is to produce standardized, internationally agreed, evidence-based datasets for use throughout the world. METHODS AND RESULTS: This article describes the development of a cancer dataset by the multidisciplinary ICCR expert panel for the reporting of thymic epithelial tumours. The dataset includes 'required' (mandatory) and 'recommended' (non-mandatory) elements, which are validated by a review of current evidence and supported by explanatory text. Seven required elements and 12 recommended elements were agreed by the international dataset authoring committee to represent the essential information for the reporting of thymic epithelial tumours. CONCLUSIONS: The use of an internationally agreed, structured pathology dataset for reporting thymic tumours provides all of the necessary information for optimal patient management, facilitates consistent and accurate data collection, and provides valuable data for research and international benchmarking. The dataset also provides a valuable resource for those countries and institutions that are not in a position to develop their own datasets.

Diffuse Idiopathic Pulmonary Neuroendocrine Cell Hyperplasia of the Lung (DIPNECH): Current Best Evidence.

Diffuse idiopathic pulmonary neuroendocrine cell hyperplasia (DIPNECH) is recognized as a preneoplastic condition by the World Health Organization. We reviewed our experience with 30 patients and performed a systematic review of the English literature to collect best evidence on the clinical features and disease course in 169 additional patients. Some patients presented with one or more carcinoid tumors associated with multiple small pulmonary nodules on imaging studies and showed DIPNECH as a somewhat unexpected pathologic finding. Others presented with multiple small pulmonary nodules that raised suspicion of metastatic disease on imaging. A third subset was presented with previously unexplained respiratory symptoms. In most patients, DIPNECH was associated with a good prognosis, with chronological progression into a subsequent carcinoid tumor noted in only one patient and death attributed directly to DIPNECH in only two patients. There is no best evidence to support the use of octreotide, steroids, or bronchodilators in DIPNECH patients.

Diagnostic difficulties with the diagnosis of small cell carcinoma of the lung.

Small cell carcinoma of the lung (SCLC) is a well characterized form of lung cancer that is frequently already metastatic at diagnosis. Thus, patients with SCLC are usually treated with chemotherapy, and therefore emphasis has been placed on distinguishing that tumor from squamous cell carcinomas, large cell carcinomas and other pulmonary neoplasms that can more often be managed surgically. SCLC can be readily and accurately diagnosed in biopsy specimens or cytological preparations, but in selected cases, it can pose difficult diagnostic dilemmas. This review discusses selected problems encountered during the pathologic diagnosis of SCLC, including its distinction from other neuroendocrine lesions such as large cell neuroendocrine carcinoma and "carcinoid" tumors." The role of immunohistology is also considered in this context.

Neuroendocrine neoplasms of the lung: Concepts and terminology.

Neuroendocrine neoplasms of the lung continue to undergo scrutiny, with respect to the diagnostic terminology recommended for them and details of their clinicopathologic profiles. This overview considers the nosological evolution of such lesions and presents current views on classification schemes that pertain to them.

Diffuse idiopathic pulmonary neuroendocrine cell hyperplasia (DIPNECH).

Diffuse idiopathic pulmonary neuroendocrine cell hyperplasia (DIPNECH) is a pre-invasive generalized proliferation of pulmonary neuroendocrine (PNE) cells, which has been described in combination with carcinoid tumorlets, obliterative bronchiolitis, and other fibrotic lung changes. Since its first recognition as a distinct syndrome in 1992, variable clinico-pathologic features have been used to diagnose DIPNECH in small case series and case reports. We recently proposed the use of the following criteria for the pathologic diagnosis of the syndrome on lung resection specimens: the presence of multifocal neuroendocrine cell hyperplasia, as defined by the presence of 5 or more PNE cells in at least 3 separate small airways combined with 3 or more carcinoid tumorlets. At diagnosis, 53% of DIPNECH patients present with a synchronous carcinoid tumor and 24% have obliterative bronchiolitis. To our knowledge, only 1 patient has been diagnosed with a carcinoid tumor subsequent to a DIPNECH diagnosis and the syndrome is not associated with an increased incidence of high-grade neuroendocrine neoplasms. Patients have usually a stable disease and only 26% experience clinical and/or radiographic deterioration. There is no evidence that treatment with octreotide or other medications improves the prognosis of DIPNECH patients.

Desmoglein 3 and p40 immunoreactivity in neoplastic and nonneoplastic thymus: a potential adjunct to help resolve selected diagnostic and staging problems.

The potential usefulness of the squamous markers p40 and desmoglein 3 (DSG-3) for the diagnosis and staging of selected thymic lesions is uncertain. We investigated their expression and distribution pattern in 66 thymomas, 12 thymic squamous carcinomas, 6 undifferentiated thymic carcinomas, 5 hyperplastic thymi, and 5 normal thymi. p40 nuclear and DSG-3 cytoplasmic/membranous immunoreactivity in greater than or equal to 10% of thymic epithelial cells was interpreted as positive, and DSG-3 distribution pattern was classified as organotypic and nonorganotypic. All nonneoplastic thymic tissues, 100% of thymic squamous carcinomas, 97% of thymomas, and 50% of undifferentiated thymic carcinomas were positive for p40. Expression of p40 in almost all thymomas and in 50% of undifferentiated carcinomas that lacked squamous features suggests that p40 is not a good marker for the diagnosis of thymic squamous carcinoma. All normal and hyperplastic thymi, 51.5% of thymomas, and 0% of thymic squamous carcinomas expressed DSG-3 in an organotypic pattern, and 13.6% of thymomas and 83% of thymic squamous carcinomas were DSG-3 positive in a nonorganotypic pattern. Findings suggest that nonorganotypic DSG-3 expression favors the diagnosis of squamous cell carcinoma over thymoma. In 26 (60.5%) of the 43 cases where neoplastic and nonneoplastic thymus were present on the same slide, the presence/absence or distribution pattern of DSG-3 immunoreactivity was different in the 2 components, suggesting that this marker can be helpful in staging thymomas with incomplete encapsulation. The presence of DSG-3-positive and DSG-3-negative thymomas raises the possibility that these tumors may originate from 2 different types of thymic epithelial cells.