Visual inspection of histopathology slides is one of the main methods used by pathologists to assess the stage, type and subtype of lung tumors. Adenocarcinoma (LUAD) and squamous cell carcinoma ...(LUSC) are the most prevalent subtypes of lung cancer, and their distinction requires visual inspection by an experienced pathologist. In this study, we trained a deep convolutional neural network (inception v3) on whole-slide images obtained from The Cancer Genome Atlas to accurately and automatically classify them into LUAD, LUSC or normal lung tissue. The performance of our method is comparable to that of pathologists, with an average area under the curve (AUC) of 0.97. Our model was validated on independent datasets of frozen tissues, formalin-fixed paraffin-embedded tissues and biopsies. Furthermore, we trained the network to predict the ten most commonly mutated genes in LUAD. We found that six of them-STK11, EGFR, FAT1, SETBP1, KRAS and TP53-can be predicted from pathology images, with AUCs from 0.733 to 0.856 as measured on a held-out population. These findings suggest that deep-learning models can assist pathologists in the detection of cancer subtype or gene mutations. Our approach can be applied to any cancer type, and the code is available at https://github.com/ncoudray/DeepPATH .
Squamous cell carcinoma arises from multiple anatomic subsites in the head and neck region. The risk factors for development of cancers of the oral cavity, oropharynx, hypopharynx, and larynx include ...tobacco exposure and alcohol dependence, and infection with oncogenic viruses is associated with cancers developing in the nasopharynx, palatine, and lingual tonsils of the oropharynx. The incidence of human papillomavirus-associated oropharyngeal cancer is increasing in developed countries, and by 2020, the annual incidence could surpass that of cervical cancer. The treatment for early-stage squamous cell cancers of the head and neck is generally single modality, either surgery or radiotherapy. The treatment for locally advanced head and neck cancers is multimodal, with either surgery followed by adjuvant radiation or chemoradiation as indicated by pathologic features or definitive chemoradiation. For recurrent disease that is not amenable to a salvage local or regional approach and for metastatic disease, chemotherapy with or without a biological agent is indicated. To date, molecular testing has not influenced treatment selection in head and neck cancer. This review will focus on the changing epidemiology, advances in diagnosis, and treatment options for squamous cell cancers of the head and neck, along with data on risk stratification specific to oropharyngeal cancer, and will highlight the direction of current trials.
We conducted a retrospective evaluation of the IMCL-9815 study to examine the association of human papillomavirus (HPV) and p16 protein expression status with outcomes in patients with oropharyngeal ...carcinoma (OPC) receiving radiotherapy (RT) plus cetuximab or RT alone.
In the IMCL-9815 study, patients were randomly allocated to receive RT plus weekly cetuximab or RT alone. A subpopulation of patients with p16-evaluable OPC was retrospectively evaluated on the basis of locoregional control (LRC), overall survival (OS), and progression-free survival (PFS). Evaluable samples from patients with p16-positive OPC were also tested for HPV DNA.
Tumor p16 status was evaluable in 182 patients with OPC enrolled in the IMCL-9815 study; 41% were p16 positive. When treated with RT alone or RT plus cetuximab, p16-positive patients had a longer OS than p16-negative patients (hazard ratio, 0.40; 95% CI, 0.21 to 0.74 and hazard ratio, 0.16; 95% CI, 0.07 to 0.36, respectively). The addition of cetuximab to RT increased LRC, OS, and PFS in both patients with p16-positive OPC and those with p16-negative disease. Interaction tests for LRC, OS, and PFS did not demonstrate any significant interaction between p16 status and treatment effect (P = .087, .085, and .253, respectively). Similar trends were observed when patients with p16-positive/HPV-positive OPC (n = 49) and those with p16-positive/HPV-negative OPC (n = 14) were compared.
p16 status was strongly prognostic for patients with OPC. The data suggest that the addition of cetuximab to RT improved clinical outcomes regardless of p16 or HPV status versus RT alone.
Oral squamous cell carcinomas (OSCC) are a heterogeneous group of cancers arising from the mucosal lining of the oral cavity. A majority of these cancers are associated with lifestyle risk habits ...including smoking, excessive alcohol consumption and betel quid chewing. Cetuximab, targeting the epidermal growth factor receptor was approved for the treatment of OSCC in 2006, and remains the only molecular targeted therapy available for OSCC. Here, we reviewed the current findings from genomic analyses of OSCC and discuss how these studies inform on the biological mechanisms underlying OSCC. Exome sequencing revealed that the significantly mutated genes are mainly tumour suppressors. Mutations in FAT1, CASP8, CDKN2A, and NOTCH1 are more frequently found in OSCC when compared to non-OSCC head and neck cancers and other squamous cell carcinomas, and HRAS and PIK3CA are the only significantly mutated oncogenes. The distribution of these mutations also differs in populations with distinct risk habits. Gene expression-based molecular classification showed that OSCC can be divided into distinct subtypes and these have a preferential response to different types of therapies, suggesting that these classifications could have clinical implications. More recently, with the approval of checkpoint inhibitors for the treatment of cancers including OSCC, genomics studies also dissected the genetic signatures of the immune compartment to delineate immune-active and -exhausted subtypes that could inform on the immune status of OSCC patients and guide the development of novel therapies to improve response to immunotherapy. Taken together, genomics studies are informing on the biology of both the epithelial and stromal compartments underlying OSCC development, and we discuss the opportunities and challenges in using these to derive clinical benefit for OSCC patients.
MiR-125a has been characterized as a tumor suppressor in several cancers. However, the role of miR-125a in cervical cancer is unknown. In this study, we found the expression of miR-125a was ...downregulated in cervical cancer patients, and negatively correlated with the tumor size, FIGO stage, and preoperative metastasis. Kaplan-Meier analysis showed that miR-125a expression predicted favorable outcome for cervical cancer patients. Dual luciferase assays identified the STAT3 gene as a novel direct target of miR-125a. Functional studies showed that miR-125a overexpression significantly suppressed the growth, invasion and epithelial-mesenchymal transition (EMT) of cervical cancer cells both in vitro and in vivo via decreasing STAT3 expression. Moreover, miR-125a conferred to G2/M cell cycle arrest, accompanied by inhibition of several G2/M checkpoint proteins. Mechanistically, inactivation of miR-125a during cervical carcinogenesis was caused by HPV suppression of p53 expression. Clinically, STAT3, the expression of which, predicted poorer outcome, was inversely correlated with miR-125a in cervical cancer. These data highlight the importance of miR-125a in the cell proliferation and progression of cervical cancer, and indicate that miR-125a may be a useful therapeutic target for cervical cancer.
BRAF V600E mutation is seen in 5% to 8% of patients with metastatic colorectal cancer (CRC) and is associated with poor prognosis. Vemurafenib, an oral BRAF V600 inhibitor, has pronounced activity in ...patients with metastatic melanoma, but its activity in patients with BRAF V600E-positive metastatic CRC was unknown.
In this multi-institutional, open-label study, patients with metastatic CRC with BRAF V600 mutations were recruited to an expansion cohort at the previously determined maximum-tolerated dose of 960 mg orally twice a day.
Twenty-one patients were enrolled, of whom 20 had received at least one prior metastatic chemotherapy regimen. Grade 3 toxicities included keratoacanthomas, rash, fatigue, and arthralgia. Of the 21 patients treated, one patient had a confirmed partial response (5%; 95% CI, 1% to 24%) and seven other patients had stable disease by RECIST criteria. Median progression-free survival was 2.1 months. Patterns of concurrent mutations, microsatellite instability status, CpG island methylation status, PTEN loss, EGFR expression, and copy number alterations were not associated with clinical benefit. In contrast to prior expectations, concurrent KRAS and NRAS mutations were detected at low allele frequency in a subset of the patients' tumors (median, 0.21% allele frequency) and were apparent mechanisms of acquired resistance in vemurafenib-sensitive patient-derived xenograft models.
In marked contrast to the results seen in patients with BRAF V600E-mutant melanoma, single-agent vemurafenib did not show meaningful clinical activity in patients with BRAF V600E mutant CRC. Combination strategies are now under development and may be informed by the presence of intratumor heterogeneity of KRAS and NRAS mutations.
The Cox proportional hazards model commonly used to evaluate prognostic variables in survival of cancer patients may be too simplistic to properly predict a cancer patient's outcome since it assumes ...that the outcome is a linear combination of covariates. In this retrospective study including 255 patients suitable for analysis who underwent surgical treatment in our department from 2000 to 2017, we applied a deep learning-based survival prediction method in oral squamous cell carcinoma (SCC) patients and validated its performance. Survival prediction using DeepSurv, a deep learning based-survival prediction algorithm, was compared with random survival forest (RSF) and the Cox proportional hazard model (CPH). DeepSurv showed the best performance among the three models, the c-index of the training and testing sets reaching 0.810 and 0.781, respectively, followed by RSF (0.770/0.764), and CPH (0.756/0.694). The performance of DeepSurv steadily improved with added features. Thus, deep learning-based survival prediction may improve prediction accuracy and guide clinicians both in choosing treatment options for better survival and in avoiding unnecessary treatments.
Treatment of orofacial tumors in dogs is associated with high morbidity and reliable prognostic factors are lacking. Dynamic contrast-enhanced computed tomography (DCECT) can be used to assess tumor ...perfusion. The objectives of this study were to describe the perfusion parameters of different types of orofacial tumors and to describe the changes in perfusion parameters during radiotherapy (RT) in a subset of them.
11 dogs with orofacial tumors prospectively recruited.
All dogs had baseline DCECT to assess blood volume (BV), blood flow (BF), and transit time (TT). Five dogs had repeat DCECT during megavoltage RT.
5 squamous cell carcinomas, 3 sarcomas, 1 melanoma, 1 histiocytic sarcoma, and 1 acanthomatous ameloblastoma were included. Blood volume and BF were higher in squamous cell carcinomas than in sarcomas, although no statistical analysis was performed. At repeat DCECT, 4 dogs showed a reduction in the size of their tumor during RT. Among these dogs, 3 showed an increase in BV and BF and 1 a decrease in these parameters between the baseline and the follow-up DCECT. The only dog whose tumor increased in size between the first and the second DCECT showed a decrease in BV and BF.
Perfusion parameters derived from DCECT were described in a series of dogs with various types of orofacial tumors. The results suggest that epithelial tumors could have higher BV and BF than mesenchymal tumors, although larger sample sizes are needed to support these preliminary findings.