How does deep learning/machine learning perform in comparison to radiologists in distinguishing glioblastomas (or grade IV astrocytomas) from primary CNS lymphomas?: a meta-analysis and systematic review.

BACKGROUND: Several studies have been published comparing deep learning (DL)/machine learning (ML) to radiologists in differentiating PCNSLs from GBMs with equivocal results. We aimed to perform this meta-analysis to evaluate the diagnostic accuracy of ML/DL versus radiologists in classifying PCNSL versus GBM using MRI. METHODOLOGY: The study was performed in accordance with PRISMA guidelines. Data was extracted and interpreted by two researchers with 12 and 23 years' experience, respectively, and QUADAS-2 tool was used for quality and risk-bias assessment. We constructed contingency tables to derive sensitivity, specificity accuracy, summary receiver operating characteristic (SROC) curve, and the area under the curve (AUC). RESULTS: Our search identified 11 studies, of which 8 satisfied our inclusion criteria and restricted the analysis in each study to reporting the model showing highest accuracy, with a total sample size of 1159 patients. The random effects model showed a pooled sensitivity of 0.89 [95% CI:0.84-0.92] for ML and 0.82 [95% CI:0.76-0.87] for radiologists. Pooled specificity was 0.88 [95% CI: 0.84-0.91] for ML and 0.90 [95% CI: 0.81-0.95] for radiologists. Pooled accuracy was 0.88 [95% CI: 0.86-0.90] for ML and 0.86 [95% CI: 0.78-0.91] for radiologists. Pooled AUC of ML was 0.94 [95% CI:0.92-0.96]and for radiologists, it was 0.90 [95% CI: 0.84-0.93]. CONCLUSIONS: MRI-based ML/DL techniques can complement radiologists to improve the accuracy of classifying GBMs from PCNSL, possibly reduce the need for a biopsy, and avoid any unwanted neurosurgical resection of a PCNSL.

Local therapy in non-metastatic primary Ewing sarcoma of the mandible and maxilla in children.

Ewing sarcoma (ES) of the jaw bones comprises a small fraction of ES at all sites. Due to their rarity, a specific policy for local treatment is lacking. The aim of this study was to evaluate the local therapy for ES and recommend measures to individualize treatment options. Patients with primary non-metastatic ES of the jaw bones treated between August 2005 and February 2015 were analyzed. All patients received primary induction chemotherapy, following which lesions amenable to resection based on specific radiological criteria were resected; those with unresectable lesions were offered definitive radiotherapy. The maxilla was the primary site in 13 patients and the mandible in eight. The median age of patients was 11.6 years (range 5-17 years). Overall, surgery was performed in 17 patients and definitive radiotherapy was used in four patients. Postoperative radiotherapy was administered to 12 patients and was avoided in five patients with 100% tumour necrosis. The 3-year overall survival, event-free survival, and local control were 68.1%, 63.6%, and 80.2%, respectively. Mandible primary and a histological response to chemotherapy were significant prognostic factors. The stratification of patients based on radiological criteria aids in selecting local therapy. In eligible patients, surgery with contemporary reconstruction results in optimal oncological and functional outcomes. Surgery also has the added advantage of identifying patients who may not need radiotherapy.

Treatment of patients with advanced gastric cancer: experience from an Indian tertiary cancer center.

Majority of patients in developing countries diagnosed with gastric cancer have an advanced stage at presentation with overall poor performance status. The aim of the study was to assess outcomes of first- and second-line chemotherapy and determine prognostic factors among patients with advanced gastric cancer (AGC). Using a prospectively maintained database, we identified 144 patients with AGC treated at Tata Memorial Centre between January 2012 and September 2013. Sixteen patients received best supportive care, and 128 patients received palliative chemotherapy. Cox regression was used for multivariate analysis of survival. Of 128 patients, 42(33%) received Cape-Ox, 22(17.1%) EOX and 47(36.7%) DOX while rest received other regimens. PS was 2 in 36 (28%) patients at presentation, and 97% of patients had ≥3 sites of metastasis. Forty-eight patients (37.5%) had signet ring histology. Median follow-up was 9 months. Median progression-free survival/overall survival (OS) was 6/8 months, respectively. Of 93 patients who progressed 39 (41.9%) patients received second-line chemotherapy. Multivariate analysis for OS showed that PS and use of taxane in first-line setting were significant prognostic factors. Patients who received second-line therapy had longer survival than those who did not (12 vs. 6 months; P=0.002). The overall outcome of our patients is comparable to the Western reported data despite an advanced disease at presentation.

Whole saliva physico-biochemical changes and quality of life in head and neck cancer patients following conventional radiation therapy: a prospective longitudinal study.

BACKGROUND: We investigated the physico-biochemical changes in saliva and its relation to quality of life (QOL) in head and neck cancer patients following conventional radiation therapy (RT). MATERIALS AND METHODS: 53 consecutive head and neck cancer patients underwent conventional RT using telecobalt photons. We analyzed objective sialometry and sialochemical parameters of salivary gland function and a physician reported Oral Assessment Protocol to assess the patients' QOL during (baseline, 3 and 6 weeks) and post RT (3 and 6 months). Statistical analysis was done using SPSS software (version 15.0; SPSS, Inc., Chicago, IL, USA). RESULTS: Stimulated salivary flow rates had shown a consistent decline during and in post-RT analysis (P < 0.001). A significant correlation was seen between mean salivary flow rates at 6 months post-RT and mean salivary electrolytes and amylase levels during the same period (P < 0.001). Mean global QOL scores had significantly worsened during RT and were still significantly poorer at 6 months than initial pre-RT levels (P < 0.001). Further, significant correlation was established between salivary pH values with global QOL scores at 6 months (P = 0.05). CONCLUSIONS: Radiation-induced hyposalivation invariably persists and correlates with poor global QOL scores seen during and following conventional RT. Post RT, there is a trend for biochemical reversal toward pre-irradiation levels suggesting a subsiding inflammation or a probable functional recovery.

Cancer of the buccal mucosa: a tale of two continents.

Squamous cell carcinoma (SCC) of the buccal mucosa in North America is thought to be different from that in the Indian subcontinent. This study compared the treatment outcomes and prognostic factors in 64 patients with SCC of the buccal mucosa treated at Cancer Care Manitoba (CCMB), Canada, and 64 similar patients treated at the Tata Memorial Hospital (TMH), India. Overall, cause-specific and disease-free survival for the two geographical groups were calculated and the impact of individual prognostic factors on survival was assessed. CCMB patients were significantly older (p<0.001), had less differentiated tumour (p=0.053) and had higher chances of positive or close surgical margins (p=0.012). At 5 years, they had lower 5-year overall survival (57.4% versus 80.1%; p<0.001), cancer-specific survival (76.4% versus 85.0%; p=0.043) and disease-free survival (42.9% versus 66.4%; p=0.004). Age had an independent influence on overall survival and cause-specific survival. After adjusting for age there was no difference in cause-specific survival between the two groups (HR=1.20; 95% CI=0.46,3.17; p=0.710). The apparent survival difference between the CCMB and TMH patients was due to the difference in the age of presentation and not because of different biological behaviour.

Cervical lymph node metastases of squamous cell carcinoma from an unknown primary: outcomes and patterns of failure.

CONTEXT: Management of cervical lymph nodes metastases of squamous cell carcinoma (SCC) from primary of unknown origin (PUO) is contentious and there is insignificant data from India on this subject. AIMS: To present experience of management of these patients treated with curative intent at a single institution. SETTINGS AND DESIGN: Retrospective study of patients treated between 1989-1994 in a tertiary referral cancer centre. MATERIALS AND METHODS: Eighty-nine patients were evaluated in the study period and their survival compared with patients with common sites of primary in the head and neck with comparable node stage. STATISTICAL ANALYSIS USED: Kaplan-Meier method. RESULTS: The clinical stage of the neck nodes at presentation was N1 in 11%, N2a in 28.5%, N2b in 22.5%, N3 in 35% and Nx in 3.4% patients. All patients underwent surgery and 70 patients received more than 40Gy postoperative radiotherapy. Twenty-nine (32.6%) patients had relapse of which 19 (21%) were in the neck. Postoperative radiotherapy did not influence the neck relapse (p=0.72). Primary was detected in 13 patients (14.6%) on subsequent follow up. The overall five and eight-years survival was 55% and 51% respectively. The overall five-year survival was better compared to patients with known primary with comparable node stage. CONCLUSIONS: Patients with cervical lymph nodes metastases of SCC from PUO have reasonable survival and low rate of development of subsequent primary when treated with surgery and radiotherapy. The overall survival is comparable to that of patients with known primary and hence an attempt at cure should always be made.

Are buccal cancers in India and Canada any different?

BACKGROUND AND OBJECTIVE: To compare the treatment outcomes of squamous cell carcinoma (SCC) of buccal mucosa in India and Canada. METHODS: We compared the outcome of 169 patients with SCC of buccal mucosa treated at Tata Memorial Hospital (TMH), India with 64 matched patients from Cancer Care Manitoba (CCMB), Canada. Overall and cause specific survivals for the two geographical groups were calculated by Kaplan-Meir method and compared using log rank test. Cox regression analysis was used to see impact of independent variables. RESULTS: At 5 years, CCMB patients had lower over all survival (57.4% vs. 67.1%; P = 0.002) than TMH ones but similar cause specific survival (76.4% vs. 74.2%; P = 0.690). Age had an independent influence on both over all and cause specific survival. After adjusting for the age confounding in the Cox proportional hazard model there was no difference in the overall survival of the two groups (HR = 0.84; 95% CI = 0.51, 1.40; P = 0.509). Radiated patients had three times higher risk of dying of disease than surgically treated ones (HR = 3.03; 95% CI = 1.64, 5.60; P < 0.001). CONCLUSIONS: There was no difference in the cause specific survival between the two groups. Apparent difference in the overall survival was due to the difference in the age of presentation.