Projection of age-standardised mortality rates for 23 cancer types in India: a nationwide projection until 2030 using autoregressive integrated moving average (ARIMA) models.

STRENGTHS AND LIMITATIONS OF THIS STUDY

This study provides the first comprehensive, long-range forecast of age-standardised mortality rates for 23 cancer types in India using a robust time-series approach.

The use of the auto.arima function with the Bayesian Information Criterion for model selection provides a principled and automated method to ensure parsimony and guard against overfitting without requiring a hold-out dataset.

The analysis is based on national-level estimates from the Global Cancer Observatory, which may mask significant subnational and regional heterogeneity in cancer patterns across states.

The accuracy of the projections is contingent on the quality and completeness of the underlying cancer mortality data, which is a known challenge in India due to the lack of a comprehensive nationwide cancer registry system.

The model does not incorporate external covariates (eg, changes in risk factor prevalence and healthcare interventions), meaning projections represent statistical extrapolations of past data and assume their continuation.

Introduction

Cancer is a global public health concern, with approximately 10 million deaths estimated in 2020, accounting for nearly 1 in 6 deaths. 1 Cancer mortality rates exhibit remarkable variations in patterns and trends across specific cancer types, and the Asian population substantially contributes to an estimated 58.3% of cancer-associated deaths, of which India is responsible for 14.7%. 1 2 In 2020, India experienced a significant cancer burden, with an estimated 851 678 cancer-related deaths and 1 324 413 newly diagnosed cancer cases, with a mortality-to-incidence ratio of 64.34%. 2

The year 2010 witnessed a substantial number of cancer-related deaths in India, with over 556 000 cases. Individuals aged 30–69 years accounted for approximately 71.1%. Notably, tobacco-related cancers caused 42% of male deaths within this group, while infections contributed to 37% of female fatalities. 3 In 2016, in India, lung, mouth, oesophagus and liver cancers significantly impacted males, while breast, cervical and lung cancers impacted females. 4 While certain cancer types experienced a decline in mortality rates, including stomach, oesophagus, leukaemia, larynx and melanoma, others showed an increase, such as lung, breast, colorectal, lymphoma, multiple myeloma, gallbladder, pancreas, kidney and mesothelioma. 5

Projections of cancer mortality would be helpful for clinicians, researchers, healthcare administrators and policymakers in planning facilities, equipment and staffing allocation, especially for regional areas. There’s also a need for such projections in health systems. Although India has a cancer registry system, there is a significant delay between case diagnosis and reporting before the data are finally entered into the registry.

Current methods to forecast cancer projections in India include linear trends and adjusted mortality-to-ratio (MI) methods. 6 7 MI has been suggested not to be used as a surrogate for cancer survival, as it is not a valid proxy for cancer survival in practice. 8 Time series analysis using autoregressive integrated moving average (ARIMA) models has been shown to have greater ability for prediction and broader applicability than non-temporal techniques. Their strength lies in the ability to explicitly model trends, cycles and other temporal structures inherent in longitudinal data. Unlike more complex ‘black-box’ machine learning models, ARIMA models offer an interpretable framework, enabling a clear understanding of the underlying data-generating process, which is crucial for policy-relevant forecasting. ARIMA delivers a robust and parsimonious baseline that has demonstrated strong performance compared with other common approaches, such as joinpoint regression. 9

Recently, ARIMA has been employed across various fields, including healthcare, to forecast data. Elsheikh et al compared the performances of a proposed long short-term memory model with those of ARIMA and Non-linear Autoregressive Artificial Neural Networks to estimate the COVID-19 transmission dynamics in Saudi Arabia. 10 A global disease burden study employed ARIMA to estimate the age-specific mortality till 2100 for 195 countries and territories. 11 Soltani et al modelled the cancer palliative care services using Seasonal ARIMA by adopting a Bayesian optimisation strategy to determine the best-performing parameters and Akaike's Information Criterion (AIC). 12

A change in the demographic profile may be seen in India, where life expectancy at birth has progressively increased from 45 years in 1971 to 63.4 years in 2002. 13 According to the Ministry of Health estimates, the life expectancy of the Indian population is projected to rise from 68 years in 2011–2015 to 71 years in 2021–2025. 14 Life expectancy among females is three points higher (72.2 years) compared with males (69.5 years) in 2020. According to the provisional population totals from the 2011 Census of India, 1429.5 million people were living in India as of 31 July 2023, up from 1028.7 million people in 2001. 14 The population will experience an increase in the incidence of non-communicable diseases like cancer, which are more likely to strike the old, as a result of these changes in age structure. Exposure to carcinogens and co-carcinogens in tobacco, insufficient exercise and, most importantly, a bad diet are the main environmental risk factors for cancer. 15

Considering the substantial impact of cancer on the Indian population, we hypothesise that predicting the number of cancer-related deaths at various cancer sites using an appropriate ARIMA model would help improve the model’s ability to forecast the age-standardised mortality rate (ASMR) of cancer across India, covering all major cancer sites, until 2030. We selected ASMR, as this metric takes into account changes in the population’s age structure over time, enabling a more accurate comparison of mortality risk across different periods, regardless of the impact of population ageing.

Materials and methods

Data source

Annual ASMR per 100 000 population for 23 cancer sites in India, covering the period 2000–2019 16 , were obtained from the Global Cancer Observatory (GCO) ‘Cancer Over Time’ database, which is maintained by the International Agency for Research on Cancer (IARC). 16 The ASMR values are standardised to the World (Segi) standard population, which allows for direct comparison of mortality risk over time by adjusting for changes in the population’s age structure. Data were extracted for males, females and both sexes combined, resulting in 69 individual time series for analysis.

Statistical analysis

Time-series forecasting was performed using ARIMA models, a class of statistical models widely used for analysing and forecasting time-series data due to their ability to capture complex temporal structures. The primary assumption for ARIMA is that the time series is stationary. The auto.arima algorithm internally tests for stationarity using unit root tests and applies the necessary degree of differencing to render the series stationary. 17 18 An ARIMA model has three components: p (order of the autoregressive component), d (degree of differencing) and q (order of the moving average component), denoted as ARIMA( p , d , q ). 17 18 The general form of the non-seasonal ARIMA model is expressed as:

Y t = Φ 0 + Φ 1 Y t − 1 + Φ 2 Y t − 2 + ⋅ ⋅ ⋅ + Φ p Y t − p + ε t − θ 1 ε t − 1 − θ 2 ε t − 2 − ⋅ ⋅ ⋅ − θ q ε t − q

d = 0 : y t = Y t

d = 1 : y t = Y t − Y t − 1

d = 2 : y t = ( Y t − Y t − 1 ) − ( Y t − 1 − Y t − 2 )

For each of the 69 time series, an optimal ARIMA( p , d , q ) model was identified using the auto.arima function from the ‘forecast’ package in R. This function implements a variation of the Hyndman-Khandakar algorithm, which automates the model selection process. The algorithm determines the appropriate degree of differencing ( d ) to make the series stationary using repeated Kwiatkowski-Phillips-Schmidt-Shin unit root tests. Further, it identifies the autoregressive ( p ) and moving average ( q ) orders that minimise the Bayesian Information Criterion (BIC).

The BIC was chosen as the selection criterion, as it fits the model imposing a penalty for each additional parameter. This adheres to the principle of parsimony and guards against overfitting by favouring simpler models unless a complex model provides a better fit to the data. The ARIMA models chosen for each site and gender combination are tabulated in online supplemental tables S1–S3 .

Model adequacy and validation

Model adequacy was assessed by examining the standardised residuals of the selected model for each time series. Standard diagnostic checks were performed, including (1) plotting the autocorrelation function of the residuals to ensure no remaining serial correlation (ie, the residuals resemble white noise) and (2) examining histograms and quantile-quantile plots of the residuals to assess their normality. 18–21 These confirmed the effective capturing of the patterns in the data, leaving residuals that were indistinguishable from noise. 18 19 To ensure robustness, the analytical code was independently verified by a second author. 18 19

Forecasting and data presentation

Once the best-fitting model for each series was identified and validated, it was used to forecast the ASMR for each year from 2020 to 2030. The projections are presented with 80% and 95% CIs to quantify the uncertainty associated with the forecasts. The results are presented in tables showing ASMR values for 2000, 2020 and the projected values for 2025 and 2030, with 95% CIs. To summarise the magnitude of change, the absolute difference in ASMR between key periods (2000–2025 and 2025–2030) was calculated. The historical data and forecasted values, along with their confidence intervals, were also visualised graphically.

Patient and public involvement

The model-based analysis employed in the study used secondary data and did not involve the public and/or patients directly in any of the phases, including planning, designing or reporting.

Results

The following results describe the projected changes in ASMR, comparing the forecasted values for 2025 and 2030 against the baseline data from 2000 ( tables 1–3 ).

The ASMR observations for both sexes (combined estimates) are projected to increase among 12 and 13 cancer sites during 2000–2025 and 2025–2030, respectively. Testicular cancer has the least slope, with a decrease in ASMR of about 33.3% and 10% during 2000–2025 and 2025–2030, respectively. Stomach cancer witnessed a reduction of 30.89% and 8.62% during 2000–2025 and 2025–2030, respectively, revealing a relative increase in mortality rate between the observation durations. It is important to note that the ASMR of the cervix, oesophageal cancer and leukaemia is shifting from positive towards negative, indicating a reduction in mortality rate. ASMR for brain, corpus uteri, melanoma and other skin cancers has remained nearly stable over time, indicating a consistent impact of these cancers during the observed period.

For males, the ASMR for all sites has experienced a decrease of 0.23% during the period from 2000 to 2025. However, it is projected to increase again by 2.77% between the years 2025 and 2030. Testicular cancer had the fewest estimates for both periods. Larynx and stomach cancer have relatively lower rates during both periods. Gallbladder, thyroid, breast, mesothelioma, prostate, brain, melanoma and other skin cancers display an increasing or stabilising estimation, as their ASMR is moving in the positive direction, considering 2000–2025 and 2025–2030. The ASMR of liver, leukaemia and oesophageal cancers also reveals a declining ASMR when we observe the 2000–2025 and 2025–2030 estimates. However, there are a few cancer sites like Colorectal Cancer (CRC), pancreas, mouth and oropharynx, kidney, lung, lymphomas and multiple myeloma cancers whose ASMR show an increasing mortality from 2000 to 2025 and 2025 to 2030. The incidence of pancreatic cancer has witnessed a substantial increase of 67.82 percentage points between the years 2000 and 2025. Moreover, it is projected to rise further by 9.59% from the period of 2025 to 2030.

For females, there were 11 and 15 sites with non-zero ASMR. Notably, cervical cancer had the highest decrease, about 27.83%, during the period from 2000 to 2025. In addition, it experienced a 5.49% reduction in ASMR during the subsequent period from 2025 to 2030. A similar projection is observed for stomach, leukaemia, oesophagus, bladder, larynx and liver, which is an alert to curtail these cancers. Contrarily, breast, trachea, bronchus, lung, CRC, pancreas, ovary, lymphomas, multiple myeloma, gallbladder, kidney and brain cancers project an increasing pattern in ASMR. Further, the graphs depicting the actual curves and forecasts have been presented in figures 1 and 2 .

Discussion

This study provides nationwide projections of mortality for 23 major cancer sites in India up to 2030. Our findings indicate a significant and diverging shift, with a projected increase in mortality from several lifestyle- and metabolically-associated cancers, with a decline in mortality from particular infection-related and tobacco-related cancers. These offer critical insights for policymakers, healthcare planners and researchers for prioritising future cancer control efforts in India.

Key findings

Our analysis projects a continued increase in the overall cancer mortality burden in India, consistent with findings from other studies that have highlighted population growth and ageing as key drivers. India, a nation of vast regional diversity, exhibits variations in cancer incidence and mortality across its states and territories. According to data from the GCO, there has been a consistent increase in the number of cancer-related deaths in India over the last two decades. 16 In 2000, the recorded deaths per 100 000 population stood at 66.36, increasing to 69.36 by 2030. The highest increase in ASMR among males was observed in CRC and pancreatic cancer during the long-term observation. The rates for CRC and pancreatic cancers increased by 29.49% and 67.81%, respectively, between 2000 and 2025 and are estimated to increase by 6.55% and 9.59%, respectively, from 2025 to 2030. Among females, breast cancer exhibits the highest increase from 2025 to 2030, followed by lung, CRC, pancreas, ovary, kidney and lymphomas, respectively. Significantly, mortality rates displayed notable variations worldwide, emerging as a leading primary or secondary cause of premature death in the age group of 30–69 years in 134 out of 183 countries. The use of ASMR is an appropriate method for accounting for changes in population demographics, specifically the age structure. Standardising the rates to a constant reference population allows for the comparison of mortality risk over time and between populations, removing the confounding effect of population ageing.

Comparison with other studies

In 45 other countries, cancer ranks third or fourth in terms of the leading cause of premature death. 22 Global mortality rates show a decline for cancers like liver and stomach, with a concerning increase in lung cancer among women. 6 Moreover, low- and middle-income countries and transitional countries are experiencing comparable mortality rates to high-income countries due to changing risk factors. This signifies an evolving landscape of challenges related to cancer across various regions and income level groups.

A recent study in India revealed that men’s cancer mortality rates decreased slightly between 2000 and 2019. 5 However, the rate exhibited a non-significant change among women. This is as observed in our study for certain cancers during the projected period. Notably, lung cancer has a 2.1% increase in annual mortality among women, in line with our ASMR, with a rise from 2.70 in 2000 to 5.02 in 2030. An increase in mouth and oropharyngeal cancer cases highlights sex-specific patterns in mortality and the variations in different types of cancers in India.

Pancreatic cancer showed the highest increase in annual mortality rate, as per Shaji et al . 5 The present study also witnessed a substantial surge for the three groups between 2000 and 2030. The projected rise can be attributed to population growth and ageing demographics. Furthermore, with the population ageing, cancer rates are likely to rise, as the disease affects older people more often. This could increase cancer-related deaths in India.

India accounts for one-third of the global oral cancer burden. 23 With 164 million users of smokeless tobacco, 69 million smokers and 42 million individuals who use both, tobacco contributes to the high incidence of oral cancer in India. 23 The southern regions of India have the highest incidence rates among women as compared with the other regions in India and abroad. 23 Tobacco, radiation, immunosuppressants, alcohol, diet, nutrition, oral thrush, Human Papillomavirus (HPV), mouthwash and dental issues contribute to oral cancer. India has a higher prevalence of advanced-stage oral cancer, with about 40% diagnosed at an advanced stage, compared with developed countries. 23–26 According to the World Cancer Report 2020, over 90% of oral cancer cases in India are among those from lower socioeconomic groups, highlighting that tobacco use is a significant public health issue. 22 Lower socioeconomic individuals are more likely to develop oral cancer because of reduced effects of interventions. 23 However, among both genders, mouth and oropharynx cancers and urinary bladder cancer mortality are projected to decrease between 2000 and 2030, attributed to the reduction in tobacco use and household air pollution. 5 High prevalence of smokeless tobacco among rural and less-educated populations leads to higher deaths among illiterate individuals than among literate ones. 3 27 Between 1990 and 2016, there was a significant decrease in the incidence of oesophageal, laryngeal and oral cancers in India. This decline can be attributed to reduced tobacco usage, which in turn lowers mortality rates associated with these types of cancers. Among all cancer sites, mouth and oropharyngeal cancers had the highest ASMR (10.21) in 2000. This rate declined slightly to 13.75 in 2030. These cancers continue to have the highest ASMR among all cancer-related deaths. However, cancer-related deaths vary across countries, with higher rates in males than in females. 1 2 15 Awareness needs to be spread among the population about the causes and fatalities of oral cancer and the importance of quitting tobacco and alcohol consumption and maintaining oral hygiene.

Between 2000 and 2025, pancreatic cancer-related deaths have witnessed a remarkable increase, with an ASMR increase of 1.72 for both sexes, 1.63 for females and 1.18 for males. There were significant increases in lymphomas, mouth and oropharynx, thyroid, CRC and kidney cancers. A notable finding is a disparity in gallbladder cancer deaths, with women experiencing higher rates than men. This may be due to sex-specific risk factors, such as reproductive history or hormonal contraceptive use. 27

In India, lifestyle-related cancers, such as breast, CRC and prostate cancers, witness an increase compared with declining infectious cancer cases. 28 29 Countries with higher levels of development have demonstrated a decline or stabilisation in incidence and mortality rates of prostate cancer, reflecting advancements in their healthcare systems and effective awareness initiatives. 6 In line with this positive behaviour, India must initiate preventive strategies to control the burden of prostate cancer effectively. 30

Breast and gynaecological cancers, including cancers of the uterus, ovaries, cervix and other reproductive organs, pose a health burden for women worldwide. Breast cancer emerged as the most commonly diagnosed cancer in 154 of 185 countries and is a major contributor to cancer-related deaths in over 100 countries. 31 All Population-Based Cancer Registries (PBCRs) in India, except Nagpur, have shown a substantial increase in the incidence of breast cancer over the years. 6 7

Public health and policy implications

The diverging trends identified in our forecasts have profound implications for India’s National Programme for Prevention and Control of Cancer, Diabetes, Cardiovascular Diseases and Stroke. The findings underscore the need for a dual-pronged strategy.

First, for cancers with steeply increasing projections, such as pancreatic, colorectal and breast cancer, there is a need to strengthen primary prevention and early detection strategies. Public health strategies should emphasise modifiable risk factors, including dietary patterns, physical activity levels and obesity. Improving early detection through expanded screening programmes, such as mammography and colonoscopy, and improving diagnostic infrastructure in rural and underserved regions are essential.

Second, for cervical and stomach cancers with declining projections, sustained efforts can maintain this positive trajectory. Scaling up HPV vaccination coverage for adolescent girls and public health measures has accelerated the decline in cervical cancers. 31 Increased cervical cancer screening programmes and awareness among women have facilitated early detection and timely treatment. 31 However, women need to continue prioritising regular screenings and vaccinations to reduce the impact of cervical cancer further. Although there is a significant increase in gastric cancer risk among younger generations in India, with possible factors contributing to the decline including reduced Helicobacter pylori infection, improved diet and decreased smoking. 32

The incidence and mortality rates of CRC worldwide are similar. In highly developed countries, there has been a noticeable decline in incidence and mortality rates, indicating successful prevention and early detection efforts. However, in India, there is a need for effective strategies to control and manage CRC. Despite shifting cancer patterns and incidence rates, India continues to face a persistent increase in overall cancer-related mortality. A higher number of breast cancer cases due to several risk factors, such as early age at menarche, late age at menopause, nulliparity, late age at first birth, fewer children, oral contraceptive use, alcohol consumption, previous history of breast cancer, smoking, genetic mutations and obesity, needs attention. 33

Although a projected stabilization is expected, the persistently high mortality rates in mouth and oropharyngeal cancers underscore the ongoing challenges of tobacco control in India. 34 The high prevalence of smokeless tobacco use continues to drive the burden of oral cancers and hinder progress. 35 Tobacco cessation programmes and enforcement of control policies need to be stringent.

Limitations of the study

Accurate recording of cancer mortality in India is hindered by several limitations, primarily due to the absence of a nationwide cancer registration system and incomplete or imprecise medical cause-of-death certifications. 6 7 While these are the best available data, their accuracy is contingent on the quality and coverage of the underlying data sources. India currently lacks a comprehensive, nationwide cancer registry system, and the GCO estimates are derived from statistical models based on data from a network of subnational PBCRs. 5 The coverage of these registries is heterogeneous across the country, which may lead to potential under-ascertainment of cases and deaths and may not fully capture the vast regional variations in cancer burden. 7 Therefore, the projections should be interpreted as national-level estimates that may mask significant state-level disparities. Discrepancies exist between statistical models and estimations, which is not unexpected, given the real-time unavailability of data. This highlights the importance of developing a robust strategy to collect accurate cancer data in the future, enabling the successful planning and better implementation of cancer control measures.

On a lower level, when the time-series models are used for predictions, the parameters of the model, that is, ( p , d , q ) in the case of ARIMA, determine the forecast estimates. Though these estimates would not broadly differ for the majority of the cases, there could be instances where there could be significant differences owing to the changes in the actual data. As a univariate time-series method, ARIMA extrapolates purely based on the historical data. The model does not incorporate external covariates, such as changes in the prevalence of risk factors, the introduction of new interventions or advancements in treatment. Consequently, the forecasts assume that the influence of these factors will continue to evolve as they have in the past. Any changes in policy, healthcare access or risk factor prevalence could lead to deviations from the projections. The projections extending 10 years are subject to uncertainty, as reflected in the widening confidence intervals. These forecasts are indications for planning purposes and not precise predictions and should be updated as new data becomes available.

Scope for future research

India is undergoing an epidemiological transition due to rapid urbanisation, industrialisation and lifestyle changes, which significantly impact the burden of cancer. This study’s findings highlight the increase and decline in cancer ASMR for 23 major cancer sites in India. These findings emphasise the importance of ongoing efforts to combat cancer and reduce mortality rates. Specific interventions can be implemented to address rising mortality rates. For instance, for cancer sites with increasing ASMR, such as CRC, pancreatic and ovarian cancers, it is crucial to focus on early detection and diagnosis and improve treatment options and access to healthcare services. Public awareness campaigns promoting healthy lifestyles and risk-reduction strategies can also play a vital role in preventing cancer and reducing mortality. Additionally, for cancer sites with declining ASMR, such as stomach and cervical cancer, continued efforts should be made to enhance preventive measures, such as widespread vaccination against HPV and initiatives to reduce risk factors, such as smoking and unhealthy dietary habits. Overall, the findings underscore the importance of a comprehensive approach that encompasses prevention, early detection, access to quality healthcare and ongoing research and development of innovative treatment options to address variations in cancer mortality rates over time effectively. Future research could compare these ARIMA-based projections with probabilistic sampling or bootstrapping to model uncertainty in the source data itself or with alternative models, such as exponential smoothing (ETS) or machine learning approaches, to perform a sensitivity analysis and further assess the robustness of the forecasts. State-specific predictions could facilitate the development of tailored interventions, subject to data availability. Hence, it is essential to implement this data collection policy to ensure proper reporting nationwide.

Conclusions

In conclusion, the findings present a nationwide projection of ASMRs for 23 cancer types in India until 2030, using ARIMA models. Despite a few limitations, our study offers valuable, data-driven projections of cancer mortality in India. The findings reveal an ongoing epidemiological transition, that is, a clear shift towards cancers associated with lifestyle and economic development. A multifaceted strategy is required to address rising rates; the best long-term approach includes raising awareness of cancer symptoms, implementing effective prevention policies, enhancing health infrastructure and deploying dedicated personnel nationwide.

Supplementary Material

Reviewer comments

Author's manuscript

Social Media: Denny John, X @djohn1976

Contributors: AS and MSN led data extraction from the sources. MSN conducted the main analysis, and VV performed the validation of the results. KP and DKV provided clinical validation. MSN led the write-up of the manuscript with inputs from AS and DJ. All authors reviewed and agreed on the final draft. DJ is the overall guarantor of the methodology and results presented in the manuscript.

Funding: Authors DJ and VV are supported with Department of Health Research, Indian Council of Medical Research, New Delhi [grant number R.12016/16/2023-HR dated 26 . 03.2024] and Indian Council of Social Science Research, New Delhi [grant number 19/LSS/2024-1054/HS&SP/SCD Dated 23.09.2024].

Competing interests: None declared.

Patient and public involvement: Patients and/or the public were not involved in the design, conduct, reporting or dissemination plans of this research.

Provenance and peer review: Not commissioned; externally peer reviewed.

Supplemental material: This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.

Data availability statement

Data are available in a public, open access repository. Supplementary tables and codes for ARIMA forecasting are available at: https://osf.io/97utz/?view_only=8d3c1fa2f29d4e8187d3a4638f208457 .

Ethics statements

Patient consent for publication

Not applicable.

Ethics approval

This study was based on publicly available, aggregated and anonymised secondary data from the GCO. No individual patient data were used. Therefore, institutional review board approval was not required for this research. As the research integrity department is not present in any of the authors' institutions, the contact detail of the Dean of Research, dean.research@msruas.ac.in , is provided.

References

1 IARC . All cancers factsheet . 2020 . Available : https://gco.iarc.fr/today/data/factsheets/cancers/39-All-cancers-fact-sheet.pdf

2 IARC. CANCER TODAY . International agency for research on cancer - world health organization . 2020 . Available : https://gco.iarc.fr/today/home

3 Dikshit R , Gupta PC , Ramasundarahettige C , et al . Cancer mortality in India: a nationally representative survey , Available : https://pubmed.ncbi.nlm.nih.gov/22460346/ 10.1016/S0140-6736(12)60358-4 22460346

4 Kulothungan V , Sathishkumar K , Leburu S , et al . Burden of cancers in India - estimates of cancer crude incidence, YLLs, YLDs and DALYs for 2021 and 2025 based on National Cancer Registry Program . BMC Cancer 2022 ; 22 : 527 . 10.1186/s12885-022-09578-1 35546232 PMC9092762

5 Shaji A , Keechilat P , DK V , et al . Analysis of the Mortality Trends of 23 Major Cancers in the Indian Population Between 2000 and 2019: A Joinpoint Regression Analysis , Available : https://pubmed.ncbi.nlm.nih.gov/36947728/ 10.1200/GO.22.00405 PMC10497286 36947728

6 Sathishkumar K , Chaturvedi M , Das P , et al . Cancer incidence estimates for 2022 & projection for 2025: Result from National Cancer Registry Programme, India . Indian J Med Res 2022 ; 156 : 598 – 607 . 10.4103/ijmr.ijmr_1821_22 36510887 PMC10231735

7 D’Souza NDR , Murthy NS , Aras RY . Projection of burden of cancer mortality for India, 2011-2026 . Asian Pac J Cancer Prev 2013 ; 14 : 4387 – 92 . 23992008

8 Ellis L , Belot A , Rachet B , et al . The Mortality-to-Incidence Ratio Is Not a Valid Proxy for Cancer Survival . J Glob Oncol 2019 ; 5 : 1 – 9 : JGO.19.00038 . 10.1200/JGO.19.00038 PMC6550058 31070980

9 Li J , Chan NB , Xue J , et al . Time series models show comparable projection performance with joinpoint regression: A comparison using historical cancer data from World Health Organization . Front Public Health 2022 ; 10 : 961449 . 10.3389/fpubh.2022.1003162 PMC9614249 36311591

10 Elsheikh AH , Saba AI , Elaziz MA , et al . Deep learning-based forecasting model for COVID-19 outbreak in Saudi Arabia . Process Saf Environ Prot 2021 ; 149 : 223 – 33 . 10.1016/j.psep.2020.10.048 33162687 PMC7604086

11 Vollset SE , Goren E , Yuan C-W , et al . Fertility, mortality, migration, and population scenarios for 195 countries and territories from 2017 to 2100: a forecasting analysis for the Global Burden of Disease Study . The Lancet 2020 ; 396 : 1285 – 306 . 10.1016/S0140-6736(20)30677-2 PMC7561721 32679112

12 Soltani M , Farahmand M , Pourghaderi AR . Machine learning-based demand forecasting in cancer palliative care home hospitalization . J Biomed Inform 2022 ; 130 : 104075 . 10.1016/j.jbi.2022.104075 35490963

13 SRS based abridged life tables, 2002-2006 . 2018 . Available : https://censusindia.gov.in/census.website/data/SRSALT

14 Provisional population totals. india: census 2011.ministry of home affairs . Government of India ; 2011 . Available : https://censusindia.gov.in/nada/index.php/catalog/42611#metadata-description

15 Bahrami H , Tafrihi M . Global trends of cancer: The role of diet, lifestyle, and environmental factors . Cancer Innov 2023 ; 2 : 290 – 301 . 10.1002/cai2.76 38089751 PMC10686168

16 IARC . The global cancer observatory - cancer over time . 2023 . Available : https://gco.iarc.fr/overtime/en

17 Prabhakaran S . ARIMA model - complete guide to time series forecasting in python . MachineLearningPlus ; 2019 . Available : https://www.machinelearningplus.com/time-series/arima-model-time-series-forecasting-python/

18 Narassima MS , Anbuudayasankar SP , Vineesh D . Dairy industry in India - impact assessment, simulation and empirical study . IJLSM 2023 ; 45 : 330 – 50 . 10.1504/IJLSM.2023.132990

19 Szmit M , Szmit A . Usage of Modified Holt-Winters Method in the Anomaly Detection of Network Traffic: Case Studies . Journal of Computer Networks and Communications 2012 ; 2012 : 1 – 5 . 10.1155/2012/192913

20 NIST. Model Validation . National institute of standards and technology, u.s . Department of Commerce ; 2023 . Available : https://www.itl.nist.gov/div898/handbook/pmc/section6/pmc624.htm

21 Gandhi P . 7 statistical tests to validate and help to fit arima model . Towards Data Science ; 2020 . Available : https://towardsdatascience.com/7-statistical-tests-to-validate-and-help-to-fit-arima-model-33c5853e2e93

22 Sankaranarayanan R , Ramadas K , Gauvreau CL , et al . Socioeconomic factors and cancer prevention in india: diverse interventions are needed . In: World Cancer Report: Cancer research for cancer prevention . Lyon (FR) : International Agency for Research on Cancer , 2020 . Available : https://www.ncbi.nlm.nih.gov/books/NBK606515/

23 Borse V , Konwar AN , Buragohain P . Oral cancer diagnosis and perspectives in India . Sensors International 2020 ; 1 : 100046 . 10.1016/j.sintl.2020.100046 34766046 PMC7515567

24 Gupta B , Bray F , Kumar N , et al . Associations between oral hygiene habits, diet, tobacco and alcohol and risk of oral cancer: A case–control study from India . Cancer Epidemiol 2017 ; 51 : 7 – 14 . 10.1016/j.canep.2017.09.003 28968558

25 Sharma S , Satyanarayana L , Asthana S , et al . Oral cancer statistics in India on the basis of first report of 29 population-based cancer registries . J Oral Maxillofac Pathol 2018 ; 22 : 18 . 10.4103/jomfp.JOMFP_113_17 29731552 PMC5917535

26 Malhotra RK , Manoharan N , Nair O , et al . Trends in Lung Cancer Incidence in Delhi, India 1988-2012: Age-Period-Cohort and Joinpoint Analyses . Asian Pac J 2018 ; 19 . Available : https://pubmed.ncbi.nlm.nih.gov/29937537/ 10.22034/APJCP.2018.19.6.1647 PMC6103588 29937537

27 Dutta U , Bush N , Kalsi D , et al . Epidemiology of gallbladder cancer in India . Chin Clin Oncol 2019 ; 8 : 33 . 10.21037/cco.2019.08.03 31484488

28 Smith RD , Mallath MK . History of the Growing Burden of Cancer in India: From Antiquity to the 21st Century . JGO 2019 ; 1 – 15 . 10.1200/JGO.19.00048 PMC7010436 31373840

29 Culp MB , Soerjomataram I , Efstathiou JA , et al . Recent Global Patterns in Prostate Cancer Incidence and Mortality Rates . Eur Urol 2020 ; 77 : 38 – 52 . 10.1016/j.eururo.2019.08.005 31493960

30 Sung H , Ferlay J , Siegel RL , et al . Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries . CA A Cancer J Clinicians 2021 ; 71 : 209 – 49 . 10.3322/caac.21660 33538338

31 Bhalerao V , Gotarkar S , Muneshwar K . The Impact of HPV Vaccination on Cervical Cancer in adolescent females: A narrative review . J Family Med Prim Care 2024 ; 13 : 4775 – 82 . 10.4103/jfmpc.jfmpc_235_24 39722908 PMC11668371

32 Kannan D , Venkatesh S , Kumar AM . Gastric Carcinoma - The Indian Perspective . Gastroenterology, Hepatology and Endoscopy Practice 2024 ; 4 : 161 – 5 . 10.4103/ghep.ghep_18_24

33 Rantala J , Seppä K , Eriksson J , et al . Incidence trends of early-onset breast cancer by lifestyle risk factors . BMC Cancer 2025 ; 25 : 326 . 10.1186/s12885-025-13730-y 39984883 PMC11846213

34 Effect of National Tobacco Control Programme (NTCP) on the clinical presentation of oral cancer in a tertiary care center of Northern India . IJPHRD 2025 ; 16 : 110 – 6 . 10.37506/xq1pp056

35 Sultana A , Krishnappa P , Verma V , et al . Association of Tobacco Use and Oral Cancer in a South Asian Population: A Case–control Study . Journal of Indian Association of Public Health Dentistry 2024 ; 22 : 355 – 61 . 10.4103/jiaphd.jiaphd_41_24