A hospital-based study of survival in oral cancer patients of Tata Memorial Hospital, Mumbai.

Introduction: Oral cancer represents a significant global public health concern, with the death rate for lip and oral cavity malignancies experiencing a 1.40-fold increase worldwide in the past three decades. This retrospective study aimed to comprehensively understand overall survival (OS) and the influence of sociodemographic and clinical factors on patients diagnosed with oral cavity cancer. Materials and methods: The study focused on oral cancer patients enrolled in 2016 and treated at Tata Memorial Hospital, Mumbai, with a follow-up period extending to 5 years until 2021. Utilising the Kaplan-Meier technique and log-rank test, we examined OS and variations based on sociodemographic factors, while the Cox proportional hazard model allowed us to investigate the simultaneous impact of multiple factors on OS. Results: A total of 1,895 eligible participants were included. The overall 5-year survival rate was 65%. After adjusting for age, gender, education, primary site, tumour grade, TNM staging, treatment intention, status and modality, we found in our study oral cancer patients aged more than 60 years (HR = 1.37, 95% CI: 1.01-1.85, p-value 0.03), patients who had poorly differentiated carcinoma (HR = 2.44, 95% CI: 1.56-3.81, p-value < 0.001), belonged to stage IV as per TNM staging (HR = 2.44, 95% CI: 1.65-3.61, p-value < 0.001), patient who have received partial treatment (HR = 2.44, 95% CI: 1.65-3.61, p-value < 0.001) and only chemotherapy (HR = 3.56, 95% CI: 2.43-5.23, p-value < 0.001) found to have a higher hazard of dying while literate (HR = 0.73, 95% CI: 0.56-0.95, p-value 0.02) are protective. Limitations: The retrospective nature of the study posed constraints in exploring additional variable associations. Implication: Overall early detection, appropriate treatment, and regular follow-up are critical for improving the survival rate of patients with oral cavity cancer. Conclusion: This research proposes that improving the socioeconomic status and promoting proactive treatment-seeking behaviour is crucial for enhancing the survival of oral cancer patients. Cancer hospitals, in collaboration with the wider public healthcare system in India, which includes clinicians and policymakers, should consider these suggestions to enhance cancer treatment and control in low-middle-income countries.

Machine Learning Diffuse Optical Tomography Using Extreme Gradient Boosting and Genetic Programming.

Diffuse optical tomography (DOT) is a non-invasive method for detecting breast cancer; however, it struggles to produce high-quality images due to the complexity of scattered light and the limitations of traditional image reconstruction algorithms. These algorithms can be affected by boundary conditions and have a low imaging accuracy, a shallow imaging depth, a long computation time, and a high signal-to-noise ratio. However, machine learning can potentially improve the performance of DOT by being better equipped to solve inverse problems, perform regression, classify medical images, and reconstruct biomedical images. In this study, we utilized a machine learning model called "XGBoost" to detect tumors in inhomogeneous breasts and applied a post-processing technique based on genetic programming to improve accuracy. The proposed algorithm was tested using simulated DOT measurements from complex inhomogeneous breasts and evaluated using the cosine similarity metrics and root mean square error loss. The results showed that the use of XGBoost and genetic programming in DOT could lead to more accurate and non-invasive detection of tumors in inhomogeneous breasts compared to traditional methods, with the reconstructed breasts having an average cosine similarity of more than 0.97 ± 0.07 and average root mean square error of around 0.1270 ± 0.0031 compared to the ground truth.

Regression-based neural network for improving image reconstruction in diffuse optical tomography.

Diffuse optical tomography (DOT) is a non-invasive imaging technique utilizing multi-scattered light at visible and infrared wavelengths to detect anomalies in tissues. However, the DOT image reconstruction is based on solving the inverse problem, which requires massive calculations and time. In this article, for the first time, to the best of our knowledge, a simple, regression-based cascaded feed-forward deep learning neural network is derived to solve the inverse problem of DOT in compressed breast geometry. The predicted data is subsequently utilized to visualize the breast tissues and their anomalies. The dataset in this study is created using a Monte-Carlo algorithm, which simulates the light propagation in the compressed breast placed inside a parallel plate source-detector geometry (forward process). The simulated DL-DOT system's performance is evaluated using the Pearson correlation coefficient (R) and the Mean squared error (MSE) metrics. Although a comparatively smaller dataset (50 nos.) is used, our simulation results show that the developed feed-forward network algorithm to solve the inverse problem delivers an increment of ∼30% over the analytical solution approach, in terms of R. Furthermore, the proposed network's MSE outperforms that of the analytical solution's MSE by a large margin revealing the robustness of the network and the adaptability of the system for potential applications in medical settings.

Imaging through diffuse media using multi-mode vortex beams and deep learning.

Optical imaging through diffuse media is a challenging issue and has attracted applications in many fields such as biomedical imaging, non-destructive testing, and computer-assisted surgery. However, light interaction with diffuse media leads to multiple scattering of the photons in the angular and spatial domain, severely degrading the image reconstruction process. In this article, a novel method to image through diffuse media using multiple modes of vortex beams and a new deep learning network named "LGDiffNet" is derived. A proof-of-concept numerical simulation is conducted using this method, and the results are experimentally verified. In this technique, the multiple modes of Gaussian and Laguerre-Gaussian beams illuminate the displayed digits dataset number, and the beams are then propagated through the diffuser before being captured on the beam profiler. Furthermore, we investigated whether imaging through diffuse media using multiple modes of vortex beams instead of Gaussian beams improves the imaging system's imaging capability and enhances the network's reconstruction ability. Our results show that illuminating the diffuser using vortex beams and employing the "LGDiffNet" network provides enhanced image reconstruction compared to existing modalities. An enhancement of ~ 1 dB, in terms of PSNR, is achieved using this method when a highly scattering diffuser of grit 220 and width 2 mm (7.11 times the mean free path) is used. No additional optimizations or reference beams were used in the imaging system, revealing the robustness of the "LGDiffNet" network and the adaptability of the imaging system for practical applications in medical imaging.

A Nomogram based prognostic score that is superior to conventional TNM staging in predicting outcome of surgically treated T4 buccal mucosa cancer: Time to think beyond TNM.

BACKGROUND AND OBJECTIVES: T4 squamous cell carcinomas of the buccal mucosa is known to have ominous outcome. The aim of this study was to develop a nomogram for T4 buccal mucosa cancer patients and demonstrate the difference in survival based on prognosticators beyond those covered by the traditional TNM staging system. METHODS: We examined medical records of treatment naïve 205 T4 buccal mucosa cancer patients operated between January 1, 2009, and December 31, 2014. A nomogram was developed using multivariate cox-regression. The nomogram was validated internally by bootstrapping and externally in an independent validation set. RESULTS: The nomogram for predicting 3-year overall survival was built using Tumor differentiation, Pathological Lymph node involvement, Bone and Perineural invasion. Based on nomogram, a score was assigned to each patient and they were divided into two groups based on Youden derived cut-off value (13.5). These two groups in training and validation set showed significant difference in survival. CONCLUSION: We developed a high performance, accurate and efficient nomogram to predict the probability of 3-year survival in T4 buccal mucosa cancer patients. Intensification of adjuvant treatment in these advanced cancer patients with poorer score might improve their survival.

Squamous cell carcinoma of buccal mucosa: An analysis of prognostic factors.

INTRODUCTION: Carcinoma of the buccal mucosa is the most common oral cavity cancer in the Indian subcontinent. The aim of this study was to analyze the outcome and evaluate prognostic factors in surgically treated buccal mucosa squamous cell carcinoma (BMSCC) patients. MATERIALS AND METHODS: A retrospective study was performed by reviewing the medical records of 409 pathologically proven buccal mucosa cancer patients, who were diagnosed and surgically treated in Tata Memorial Hospital between January 1, 2006, and December 31, 2008. RESULTS: The overall 5-year survival of the cohort was found to be 54.1%. The stage-wise survival rate for tumor, node, metastasis (TNM) Stage I, II, III, and IV patients was found to be 85.2%, 82.9%, 56.3%, and 42.6% (P < 0.00), respectively. On multivariate Cox proportional hazard analysis, the presence of comorbidity, histological tumor size, pathological lymph node status, tumor differentiation, perineural invasion, and extracapsular spread were found to be independently associated with overall survival. CONCLUSION: BMSCC is an aggressive malignant tumor. In addition to TNM classification, other clinical and pathological factors also have a significant role in BMSCC prognostication. Hence, there is a need to move beyond TNM and develop a more inclusive, flexible, and easy to use prognostic system.

Sociodemographic Factors and Late-stage Diagnosis of Breast Cancer in India: A Hospital-based Study.

CONTEXT: Breast cancer (BC) is one of the major causes of cancer mortality in India. Late-stage diagnosis of BC is associated with poor survival. Identification of factors affecting late presentation of the disease could be an effective step to reduce BC mortality. AIMS: To study the association of sociodemographic factors with BC stage at diagnosis. SETTINGS AND DESIGN: The study is a retrospective analysis from the case records from a single institution. SUBJECTS AND METHODS: Data for the year 2008 was collected from the hospital records. A total of 1210 cases were included for the analysis. Sociodemographic factors included were age, place of residence, religion, marital status, level of education, and occupation. Other study variables were family history, presence of comorbidity, and stage at diagnosis. STATISTICAL ANALYSIS: Association between sociodemographic factors by stage at diagnosis was tested using Chi-square statistics, with odds ratios (ORs) estimated through logistic regression modeling. RESULTS: In the study cohort, 46% patients had reported at early stages and 54% at advanced stages. All factors were evaluated for being predictors of disease stage at presentation using univariate and multivariate logistic regression model. Women from urban background were less likely to present with advanced stage disease (OR = 0.64; 95% confidence interval [CI]: 0.49-0.84) as compared to rural women. Similarly, illiterate women were also more likely to present with advanced-stage disease (OR = 1.55; 95% CI: 1.16-2.09). CONCLUSIONS: This data clearly indicate that the patients of rural background and of low education status are more likely than their respective counterparts to have an advanced stage of BC diagnosis. Our results may be considered the keys to determining how stage variation may be related to patients and community characteristics and where limited resources need to be invested to ensure early diagnosis of BC.

Nomogram prediction for survival of patients with oral cavity squamous cell carcinoma.

BACKGROUND: Nomograms are visual predictive tools that are widely used for estimating cancer prognosis. The purpose of this study was to develop a prognostic tool that will accurately predict overall survival (OS) in individual patients with oral cancer. METHODS: Medical records of 609 patients with cancer of the oral cavity were retrospectively analyzed. The stepdown method was used to select the statistically most influential predictors for inclusion in the final nomogram for predicting OS. RESULTS: The most influential predictors of mortality probability were age, comorbidities, clinical lymph node status, stage of disease, tumor thickness, differentiation, and perineural invasion. Nomogram was developed for prediction of 5-year OS. CONCLUSION: Nomogram has been widely used in a number of cancers, but its application in oral cancer prognostication has not been explored. In this study, we have developed a nomogram that can accurately estimate 5-year OS in patients with oral cavity cancer. © 2016 Wiley Periodicals, Inc. Head Neck 38: 1826-1831, 2016.

Burden of cervical cancer and role of screening in India.

BACKGROUND: Cervical cancer is a major cause of cancer mortality in women and more than a quarter of its global burden is contributed by developing countries. In India, in spite of alarmingly high figures, there is no nationwide government-sponsored screening program. This study was conducted to assess the burden of cervical cancer in India and review the performance characteristics of available cervical cancer screening tools, so as to provide evidence-based recommendations for application of most practically suited screening test to be used in resource-poor field settings. MATERIALS AND METHODS: MEDLINE and Web of Science electronic database were searched from January 1990 to December 2015, using the keywords such as "cervical cancer", "screening", "early detection", "cervical cytology" and "visual inspection", and their corresponding MeSH terms in combination with Boolean operators "OR, AND." Two authors independently selected studies that are published in English and conducted in India. A total of 11 studies were found to be relevant and eligible to be included in the present study. RESULTS: In India, cervical cancer contributes to approximately 6-29% of all cancers in women. The age-adjusted incidence rate of cervical cancer varies widely among registries; highest is 23.07/100,000 in Mizoram state and the lowest is 4.91/100,000 in Dibrugarh district. The pooled estimates of sensitivity and specificity of visual inspection with acetic acid (VIA), magnified VIA, visual inspection with Lugol's iodine (VILI), cytology (Pap smear), and human papillomavirus DNA were found to be 67.65% and 84.32%, 65.36% and 85.76%, 78.27% and 87.10%, 62.11% and 93.51%, and 77.81% and 91.54%, respectively. CONCLUSIONS: In developing countries because of lack of necessary infrastructure and quality control, high-quality cytology screening may not be feasible for wide-scale implementation. Hence, cervical cancer screening program based on visual screening test such as VIA/VILI should be adopted as an integral part of primary health-care setup in resource-poor countries like India.

Lifestyle factors including diet and leukemia development: a case-control study from Mumbai, India.

In India, among males, leukemia rates vary across the country. The present unmatched hospital-based case- control study conducted at Tata Memorial Hospital included subjects registered between the years 1997-99. There were 246 leukemia cases and 1,383 normal controls. Data on demographics, lifestyle, diet and occupation history were recorded. Cigarette (OR=2.1) and bidi smoking (OR=3.4) showed excess risk for leukemia. Odds ratios were 3.9 for fish-eaters, 0.40 for chilli eaters, 1.5 for milk drinkers and 0.60 for coffee drinkers, compared to non-drinkers/eaters. However, neither exposure to use of pesticides nor cotton dust showed any excess risk for leukemia.

Prostate cancer: a hospital-based survival study from Mumbai, India.

BACKGROUND: Prostate cancer is common in elderly men, especially in western countries, and incidences are rising in low-risk populations as well. In India, the age-standardized rates vary between registries. Under these circumstances we have estimated the survival of prostate cancer patients based on age, family history, diabetes, hypertension, tobacco habit, clinical extent of disease (risk group) and treatment received. MATERIALS AND METHODS: The present retrospective study was carried out at the Tata Memorial Hospital (TMH), Mumbai, India. During years 1999-2002, some 850 prostate cancer cases, including 371 new cases, treated in TMH were considered as eligible entrants for the study. Five-year survival rates using actuarial and loss-adjusted (LAR) method were estimated. RESULTS: The patient population was distributed uniformly over the three age groups. A larger proportion of the patients were diagnosed at 'metastatic stage' and hormone treatment was most common. 20% patients had history of diabetes and 40% with hypertension. The 5-year overall survival rate was 64%. Survival was 55%, 74% and 52% for ' <59 years', '60-69 years' and ' >70 years' respectively. Non-diabetic (70%), hypertensive (74%), with family history (80%) of cancer, with localized-disease (91%) and treated with surgery, either alone or in combination, (91%) had better survival. CONCLUSIONS: The present study showed that prostate cancer patients with localized disease at diagnosis experience a better outcome. Local treatment with either surgery or radiation achieves a reasonable outcome in prostate cancer patients. A detailed study will help in understanding the prognostic indicators for survival especially with the newer treatment technologies available now.

Case-control study of risk factors for Non-Hodgkin lymphoma in Mumbai, India.

BACKGROUND: In the year 2010, it is estimated that nearly 0.36 million new cases and 0.19 million deaths with Non-Hodgkin lymphoma occurred. In India, among males, NHL incidence rates vary across the country which has encouraged us to conduct a case-control study to study risk factors. MATERIALS AND METHODS: The present unmatched hospital-based case-control study conducted at Tata Memorial Hospital included subjects registered between the years 1997-99. There were 390 'lymphoma cases' and 1,383 'normal controls. RESULTS: Data on age, tobacco habits, occupational history, dietary factors, tea, coffee were collected by the social investigators. Univariate and multivariate methods were applied for obtaining the odds ratios for risk factors. CONCLUSIONS: In the study, cigarette smoking (OR=2.0) and bidi smoking (OR=2.8), were associated with excess risk of lymphoma. Among the dietary items, only consumption of mutton showed 7.3-fold significant excess risk for lymphoma. Consumption of milk showed a 6-fold excess risk (OR=1.5); while coffee showed a 50% reduction in risk for lymphoma. Among occupational exposure, exposure to use of pesticides showed 3-fold excess risk for lymphoma.

Hospital-based study of endometrial cancer survival in Mumbai, India.

BACKGROUND: Endometrial cancer is common in western women, and the rates are very high; however in India, the rates are as low as 4.3 per 100,000 (Delhi). OBJECTIVE: To estimate the survival of endometrial cancer patients based on age, education, family history, tobacco habit, number of pregnancies, clinical extent of disease and treatment received. MATERIALS AND METHODS: The present retrospective study was carried out at the Tata Memorial Hospital (TMH), Mumbai, India, between 1999-2002. 310 cases treated in TMH were considered as eligible entrants for the study. Five-year survival rates were estimated using actuarial and loss-adjusted (LAR) methods. RESULTS: The proportions of patients dying above 50 years of age, non-residents and illiterates was higher than their counterparts. 54.8% of patients had some form of treatment before attending TMH. There were only 4.2% tobacco-chewers and only 6.1% had a family history of cancer. There were 25.8% who had 3-5 pregnancies (not living children) and 38.1% did not remember the pregnancy history. The 5-year overall survival rate was 92%. The five-year rates indicated better prognosis for those aged less than 50 years (97%), non-tobacco-chewers (94%), with no family history of cancer (93%), with localized disease (93%) and those treated with surgery either alone or as a combination treatment (95%). CONCLUSIONS: The present study showed that endometrial cancer patients with localized disease at diagnosis have a good outcome in India. A detailed study will help in understanding the prognostic indicators for survival especially with the newer treatment technologies now available.