Spectrum of Nasopharyngeal Cancers Seen in the Department of Radiation Oncology University College Hospital, Ibadan

Background: Nasopharyngeal cancer was initially thought to be an uncommon disease in Nigeria, but recent studies have shown a steady increase in the incidence of the disease with an associated increase in morbidity and mortality due to late presentation. Method: This was a retrospective review of patients with nasopharyngeal cancer seen in the department of Radiation Oncology U.C.H. Ibadan from 2000 to 2009. Patient's bio-data, clinical-pathologic presentation and site of referral were retrieved using data extraction form and subsequently analysed using Micrsoft Exel (2013). Results: A total of 205 patients with histologically confirmed nasopharyngeal cancer seen during the study period were analysed. The mean age of the patients was 41.7 years with 144 (70.2%) males and 61(29.8%) females. The commonest presenting complaint was cervical lymphadenopathy in 98.5% of patients followed by nasal blockade, nasal discharge and epistaxis with 82.0%, 78.0% and 51.2% respectively. Undifferentiated carcinoma was the commonest histologic type (68.8%). Majority of patients were civil servants (33.2%). The habit of smoking and alcohol ingestion were unspecified in majority of patients (43.4%). Ingestion of smoked fish has been found in majority of patients 122(59.5%) with the disease in this study. Most of the patients came from southern part of the country which coincide with geographic distribution of Burkitt lymphoma, hence the possibility of Ebsteinbarr virus (EBV) as one of the likely causative agents. Conclusion: Nasopharyngeal cancer is common. Cervical lymphadenopathy, nasal blockade with undifferentiated histology and late presentation were the norm. Referrals were from all the regions of the country.

Small Cell Undifferentiated Histology Does Not Adversely Affect Outcome in Hepatoblastoma: A Report From the Children's Oncology Group (COG) AHEP0731 Study Committee

PURPOSE Small cell undifferentiated (SCU) histology in hepatoblastoma (HB) tumors has historically been associated with a poor prognosis. Tumors from patients enrolled on Children's Oncology Group (COG) study AHEP0731 underwent institutional and central pathologic review for identification of SCU histology. PATIENTS AND METHODS Patients with SCU histology identified at the local treating institution who had otherwise low-risk tumors were upstaged to the intermediate-risk treatment stratum, whereas those only identified by retrospective central review were treated per the local institution as low-risk. Patients with otherwise intermediate- or high-risk tumors remained in that treatment stratum, respectively. Central review was to be performed for all tissue samples obtained at any time point. Treatment was per local review, whereas analysis of outcome was based on central review. RESULTS Thirty-five patients had some elements (1%-25%) of SCU identified on central review of diagnostic specimens. All but two patient tissue sample retained nuclear INI1 expression. The presence of SCU histology did not correlate with age, alpha-fetoprotein level at diagnosis, or sex. The presence of SCU did not affect event-free survival (EFS). EFS at 5 years for patients with low-risk, intermediate-risk, and high-risk with SCU HB was 86% (95% CI, 33 to 98), 81% (95% CI, 57 to 92), and 29% (95% CI, 4 to 61), respectively, compared with EFS at 5 years for patients without SCU enrolled with low-risk, intermediate-risk, and high-risk of 87% (95% CI, 72 to 95), 88% (95% CI, 79 to 94), and 55% (95% CI, 32 to 74; P = .17), respectively. CONCLUSION The presence of SCU histology in HB does not appear to adversely affect outcome. Future studies should be able to treat patients with SCU HB according to risk stratification without regard to the presence of SCU histology.

Expanding the Immunophenotype Spectrum of SMARCA4-Deficient Non-Small Cell Lung Carcinomas: A Case Series with Neuroendocrine Markers Expression

Aims. In recent years, SMARCA4-deficient nonsmall cell lung cancer (NSCLC) has been recognized as a distinct new subtype of lung cancer, which is characterized by loss of SMARCA4 (Brahma-related gene-1 [BRG1]) protein expression. Only a limited number of SMARCA4-deficient NSCLC case series have been reported, and their clinicopathological features have not yet been fully elucidated. Our main aim was to analyze the clinical history, histology, immunohistochemistry, and molecular pathology of 5 SMARCA4-deficient NSCLC patients with poorly differentiated or undifferentiated histology and neuroendocrine markers expression. Methods and results. Five patients with complete loss of nuclear BRG1 immunostaining were identified among 53 patients of poorly differentiated/undifferentiated NSCLC. We then performed immunohistochemical staining and gene mutation analysis using a real-time polymerase chain reaction. All patients were male aged between 58 and 82 years (average 67.6 years), with smoking exposure. Histologically, the tumors had a relatively monotonous morphology and showed solid nest-like, sheet-like growth, and geographic necrosis. Thyroid transcription factor 1, cytokeratin 7, and Napsin A were all negative (5 of 5). Moreover, all tumors showed a variable expression of neuroendocrine markers, including synaptophysin, chromogranin A and CD56. Hot spot epidermal growth factor receptor/anaplastic large-cell lymphoma kinase/c-ros oncogene 1 mutations were not detected in any of the 5 tumors. Conclusions. To the best of our knowledge, this is the first study that has reported the poorly differentiated morphology with a frequent expression of neuroendocrine markers. Our results have expanded the immunophenotype spectrum of SMARCA4-deficient NSCLC. However, the clinicopathological significance of this subset of SMARCA4-deficient NSCLC should be further clarified in larger series studies.

High-Risk Endometrial Cancer Assessed by Immediate Intraoperative Frozen Sections of Sentinel Lymph Nodes - A Retrospective Study.

While sentinel lymph node (SLN) sampling has been established for low-risk endometrial cancer, few data exists on high-risk histologies. This study aims to measure the accuracy of immediate intraoperative SLN biopsy with frozen section in high-risk endometrial cancer. Patients diagnosed with endometrial cancer of high-grade histology (grade 3 endometrioid, clear cell, serous, carcinosarcoma, de- or undifferentiated histology) between 2014 and 2019 at a single institution who underwent SLN mapping, followed by pelvic lymphadenectomy with or without para-aortic lymphadenectomy were included. SLNs were assessed intraoperatively using multiple frozen sections and H&E staining. Lymph node metastases detected by SLN biopsy were compared with complete lymphadenectomy specimens. 35 patients with high-grade endometrial cancer histology underwent SLN mapping followed by lymphadenectomy. In 34 of 35 (97%) of these patients mapping with at least one SLN was successful. Positive SLNs were identified in 7/34 patients (20.6%). There were no patients who had positive lymph nodes on complete lymphadenectomy without a positive SLN, resulting in 100% sensitivity, and 0% false-negative rate. SLN mapping using intraoperative frozen sections in high-risk endometrial cancer demonstrated 100% sensitivity and 0% false-negative rate, provides immediate feedback on successful SLN mapping and valuable intraoperative information on the disease status guiding the intraoperative decision for completion lymphadenectomy.

Establishing Machine Learning Models to Predict Curative Resection in Early Gastric Cancer with Undifferentiated Histology: Development and Usability Study

Background Undifferentiated type of early gastric cancer (U-EGC) is included among the expanded indications of endoscopic submucosal dissection (ESD); however, the rate of curative resection remains unsatisfactory. Endoscopists predict the probability of curative resection by considering the size and shape of the lesion and whether ulcers are present or not. The location of the lesion, indicating the likely technical difficulty, is also considered. Objective The aim of this study was to establish machine learning (ML) models to better predict the possibility of curative resection in U-EGC prior to ESD. Methods A nationwide cohort of 2703 U-EGCs treated by ESD or surgery were adopted for the training and internal validation cohorts. Separately, an independent data set of the Korean ESD registry (n=275) and an Asan medical center data set (n=127) treated by ESD were chosen for external validation. Eighteen ML classifiers were selected to establish prediction models of curative resection with the following variables: age; sex; location, size, and shape of the lesion; and whether ulcers were present or not. Results Among the 18 models, the extreme gradient boosting classifier showed the best performance (internal validation accuracy 93.4%, 95% CI 90.4%-96.4%; precision 92.6%, 95% CI 89.5%-95.7%; recall 99.0%, 95% CI 97.8%-99.9%; and F1 score 95.7%, 95% CI 93.3%-98.1%). Attempts at external validation showed substantial accuracy (first external validation 81.5%, 95% CI 76.9%-86.1% and second external validation 89.8%, 95% CI 84.5%-95.1%). Lesion size was the most important feature in each explainable artificial intelligence analysis. Conclusions We established an ML model capable of accurately predicting the curative resection of U-EGC before ESD by considering the morphological and ecological characteristics of the lesions.

Undifferentiated large cell/rhabdoid carcinoma presenting in the intestines of patients with concurrent or recent non-small cell lung cancer (NSCLC): clinicopathologic and molecular analysis of 14 cases indicates an unusual pattern of dedifferentiated metastases

Undifferentiated carcinoma metastatic to the bowel is uncommon in surgical pathology practice and might be confused with primary gastrointestinal carcinoma, melanoma, lymphoma, and others. We present 14 cases of uni- (n = 9) or multifocal (n = 5) undifferentiated large cell/rhabdoid carcinoma presenting in the bowel of patients with concurrent (n = 9) or recent (diagnosed 1 to 25 months earlier; median, 4) non-small cell lung cancer (NSCLC). Patients were 6 females and 8 males, aged 52 to 85 years. Primary NSCLC was verified histologically in 10 cases and by imaging in 4. The undifferentiated histology was present in the lung biopsy in 4/10 patients (as sole pattern in 3 and combined with adenocarcinoma in 1) and was limited to the intestinal metastases in the remainder. PDL1 was strongly expressed in 7/9 cases (CPS: 41 to 100). Loss of at least one SWI/SNF subunit was detected in 7/13 cases (54%). SMARCA2 loss (n = 6) was most frequent and was combined with SMARCA4 loss in one case. PBRM1 loss was observed in one tumor. Successful molecular testing of 11 cases revealed BRAF mutations in 4 (3 were non-V600E variants), KRAS mutations in 3, and wildtype in 4. None had EGFR mutations. Analysis of 4 paired samples revealed concordant KRAS (2) and BRAF (1) mutations or wildtype (1). Our study indicates that undifferentiated carcinoma within the intestines of patients with concurrent/recent NSCLC represents dedifferentiated metastasis from the NSCLC. Recognition of this unusual presentation is cardinal to avoid misdiagnosis with inappropriate therapeutic and prognostic implications.

Advanced T Stage is the Sole Determining Factor of Survival in Patients with Nasopharyngeal Carcinoma: The Outcomes of the Multi-Modality Management

Background: Thailand is an endemic area of nasopharyngeal carcinoma (NPC). Multi-modality treatment results in significant improvement in survivals. Objective: The overall survival (OS) of patients with early and locally-advanced (E and LA, stage I-IVa and IVb) diseases was the primary objective. Materials and Methods: The present study was a retrospective cohort study of patients with NPC treated at Vajira Hospital between 2013 and 2016. Baseline characteristics including age, gender, histopathology, staging, modality of treatment, time to radiotherapy completion, serious adverse events, treatment responses, patterns of recurrence, and metastasis were collected. Results: One hundred patients with mostly undifferentiated histology that presented with LA disease were included in this study. Stage III, IVa, and IVb accounted for 28%, 18%, and 22%, respectively. After median follow-up of 41.1 months, the median OS of patients E/LA diseases (stage I, II, III, IVa, and IVb) was not reached. Neither induction chemotherapy (IC) nor adjuvant chemotherapy (AC) was associated with superior disease-free survival (DFS) compared to definitive concurrent chemoradiotherapy (CRT) alone across all stage subgroups. Only T3 and T4 disease were significantly related to worse OS. Conclusion: With the standard CCRT, patients with E and LA diseases had excellent survival outcomes compared to the results from international studies. Keywords: Nasopharyngeal carcinoma, Multi-modality treatment, Outcomes, Real-life practice

A Comprehensive Analysis of Treatment Management and Survival Outcomes in Nasopharyngeal Carcinoma

Objectives To comprehensively investigate nasopharyngeal carcinoma (NPC) treatment, overall survival (OS), and the influence of clinical/sociodemographic factors on outcome. Study Design Retrospective database study. Setting National Cancer Database. Methods The 2004-2015 National Cancer Database was queried for all patients with NPC receiving definitive treatment. Log-rank tests and Cox proportional hazards models were used for statistical analyses. Results A total of 8260 patients with NPC were included (71.4% male; 42.5% with keratinizing histology; mean ± SD age, 52.1 ± 15.1 years), with a 5-year OS of 63.4%. Multivariate predictors of mortality included age ≥65 years (hazard ratio [HR], 1.81; P < .001), Charlson/Deyo score ≥1 (HR, 1.27; P = .001), American Joint Committee on Cancer clinical stage III to IV (HR, 1.85; P < .001), and government insurance or no insurance (HR, 1.53; P < .001). Predictors of survival included female sex (HR, 0.82; P = .002), Asian/Pacific Islander race (HR, 0.74; P < .001), nonkeratinizing/undifferentiated histology (HR, 0.79; P = .004), and receiving treatment at academic centers (HR, 0.87; P = .02). Chemoradiotherapy (CRT) demonstrated improved OS as compared with radiotherapy (RT) only for stage II ( P = .006) and stage III ( P = .005) and with RT or chemotherapy only in stage IVA NPC ( P < .001). When compared with CRT alone, surgery plus CRT provided OS benefits in keratinizing ( P = .013) or stage IVA ( P = .030) NPC. When compared with RT, CRT provided OS benefits in keratinizing ( P = .005) but not nonkeratinizing ( P = .240) or undifferentiated ( P = .390) NPC. Substandard radiation dosing of <60 Gy and <30 fractions were associated with inferior OS (both P < .001). Conclusions NPC survival is dependent on a variety of clinical/sociodemographic factors. Stage-specific treatments with optimal OS include CRT or RT for stages I to II and CRT for stage III to IV. The large representation of nonendemic histology is valuable, as these cases are not well characterized.

Establishing Machine Learning Models to Predict Curative Resection in Early Gastric Cancer with Undifferentiated Histology: Development and Usability Study (Preprint)

BACKGROUND Undifferentiated type of early gastric cancer (U-EGC) is included among the expanded indications of endoscopic submucosal dissection (ESD); however, the rate of curative resection remains unsatisfactory. Endoscopists predict the probability of curative resection by considering the size and shape of the lesion and whether ulcers are present or not. The location of the lesion, indicating the likely technical difficulty, is also considered. OBJECTIVE The aim of this study was to establish machine learning (ML) models to better predict the possibility of curative resection in U-EGC prior to ESD. METHODS A nationwide cohort of 2703 U-EGCs treated by ESD or surgery were adopted for the training and internal validation cohorts. Separately, an independent data set of the Korean ESD registry (n=275) and an Asan medical center data set (n=127) treated by ESD were chosen for external validation. Eighteen ML classifiers were selected to establish prediction models of curative resection with the following variables: age; sex; location, size, and shape of the lesion; and whether ulcers were present or not. RESULTS Among the 18 models, the extreme gradient boosting classifier showed the best performance (internal validation accuracy 93.4%, 95% CI 90.4%-96.4%; precision 92.6%, 95% CI 89.5%-95.7%; recall 99.0%, 95% CI 97.8%-99.9%; and F1 score 95.7%, 95% CI 93.3%-98.1%). Attempts at external validation showed substantial accuracy (first external validation 81.5%, 95% CI 76.9%-86.1% and second external validation 89.8%, 95% CI 84.5%-95.1%). Lesion size was the most important feature in each explainable artificial intelligence analysis. CONCLUSIONS We established an ML model capable of accurately predicting the curative resection of U-EGC before ESD by considering the morphological and ecological characteristics of the lesions.