|Ahead of print
Incidence of cancers of unknown primary origin in India and their trends during 1986–2014: A joinpoint regression analysis
Satyanarayana Labani, Smita Asthana, Parul Vohra, Uma Kailash, Abhinav Srivastava
Division of Epidemiology and Biostatistics, National Institute of Cancer Prevention and Research, Indian Council of Medical Research, Noida, Uttar Pradesh, India
|Date of Submission||05-Sep-2019|
|Date of Decision||03-Feb-2020|
|Date of Acceptance||17-Feb-2020|
|Date of Web Publication||21-Mar-2021|
Division of Epidemiology and Biostatistics, National Institute of Cancer Prevention and Research, Indian Council of Medical Research, Noida, Uttar Pradesh
Source of Support: None, Conflict of Interest: None
Background: Cancer of unknown primary (CUP) origin is cancer in which malignant cancer cells are in the body but the site of cancer where it began is unknown. Detailed incidence and time trends of these cancers, specific to various regions in India is needed. This paper aims to summarize and report the incidence of other and unknown (O&U) cancers across India in 27 population-based cancer registries (PBCRs) and to study the trends among these cancers using joinpoint regression analysis.
Methods: Data on the incidence of CUP were obtained from the published reports on 27 PBCRs of the National Cancer Registry Programme (NCRP) of the Indian Council of Medical Research (ICMR). A joinpoint regression model was used to analyze the long-term trends of incidence related to CUP based on published data from PBCRs between 1986 and 2014. Annual Percent Change (APC) in incidence rate was estimated for various registries.
Results: The northeast region had the highest age-adjusted rate (AAR) for both men (1.76–29.7) and women (1.99–14.68). Age-specific rate (ASR) for both men (39.8–855.7) and women (48.2–470.4) was highest in the northeast region. There is an increase in the incidence rate for all six major registries over the past decade with an exception of women in the Delhi Cancer Registry. There is a decline in incidence rate by 0.14 during 1990–2012 in the female population of the Delhi registry.
Conclusion: The increasing incidence trends of CUP is a matter of concern for the healthcare professionals and researchers. There is a need for research and advanced and improved diagnostic tools for the improvement of the status of O&U cancers.
Keywords: Joinpoint regression analysis, PBCR, trends, unknown primary neoplasms
Key Message The increasing incidence rate of other and unknown cancers is a matter of concern. Detailed research and improved and advanced diagnostic tools are required for better treatment.
|How to cite this URL:|
Labani S, Asthana S, Vohra P, Kailash U, Srivastava A. Incidence of cancers of unknown primary origin in India and their trends during 1986–2014: A joinpoint regression analysis. Indian J Cancer [Epub ahead of print] [cited 2022 Sep 28]. Available from: https://www.indianjcancer.com/preprintarticle.asp?id=311636
| » Introduction|| |
In the majority of advanced cancer cases, the organ site of primary lesion becomes shortly evident after clinical, pathological, radiological, and other evaluations. The International Classification of Diseases (ICD) recognizes that a certain proportion of cancers are either found to be secondary in the organ biopsied or of the unknown primary site. Carcinoma of unknown primary (CUP) origin is a diverse group of cancers defined by the presence of metastatic disease where the primary cancer site remains unidentified. The symptoms and signs of CUP differs, depending on the spread of cancer in the body. Diagnosis of CUP requires a clinical picture consistent with metastatic disease and one or more biopsy results inconsistent with a primary tumor. Tests and procedures used to find primary cancer depend on the spread of cancer. Thus, the unknown primary tumor is defined as biopsy-proven metastasis of a malignancy in the absence of an identifiable primary site after complete history, physical examination, basic laboratory studies, chest X-ray, and additional directed studies indicated by positive findings during the initial workup.
The incidence of CUPs is about 3–5% of all diagnosed cancers and a poor prognosis in most (80–85%) of the cases. The other 15–20% of patients, however, have a relatively long survival with appropriate treatment. The annual age-adjusted incidence of CUP is highest in Australia (18–19) per 100,000 populations. There were around 8,800 new cancers of unknown primary cases in the UK every year for years 2014–2016. A study conducted in Guwahati, India from January 2010 to December 2011 revealed that out of 12,285 patients 554 had CUP which accounts for 4.5%. Although this malignant syndrome accounts for 3–5% of all cancer diagnosis, the majority of patients still lack effective therapeutic regimens.
Cancer registries provide statistical data on various cancers including unknown primary cancers. A cancer registry is an organization for systematic collection, storage, analysis, interpretation, and reporting of data on subjects with cancer. Population-based cancer registries (PBCRs) are primarily concerned with estimating cancer burden in a population residing in a defined geographic region. The data provide a description of cancer incidence and mortality, which helps in the planning and evaluation of cancer control activities. There are some region-specific studies from India to calculate the burden of others and unknown (O&U) cancers., There is no study available on the incidence and trend of these cancers in the entire Indian context. Data on O&U cancers available in various caner registries need summarization which will help researchers and policymakers to understand the burden of these cancers. Thus, we aimed to summarize and report current incidence and trends, and O&U cancers using data from the National Cancer Registry Programme (NCRP) for selected registries.
| » Methodology|| |
Data collection and reporting
Data collected on the incidence of O&U cancers for the present study are obtained from the published consolidated reports on 27 PBCRs for 2012–2014 of NCRP of the Indian Council of Medical Research (ICMR). The PBCRs in various parts across the country were divided into six regions for the purpose of the present study which are as follows: North: Delhi; Patiala, South: Bangalore; Chennai; Kolla; Thiruvananthapuram, Central: Bhopal, East: Kolkata, Northeast: Cachar district; Kamrup urban; Manipur; Mizoram; Nagaland; Meghalaya; Sikkim; Tripura, West: Mumbai; Nagpur; Pune; Ahmedabad and Barshi extended. The data of Indian registries are available for public use on the NCRP-ICMR website.
This data from NCRP were also available on the broad cancer category, that is, O&U cancers. As per ICD-10 codes, O&U cancers include C26, C39, C48, C75, C76, C77, C78, C79, C80, and C97. For O&U cancers, age-adjusted rate (AAR) up to age 74 years and age-specific rates (ASRs) up to age 65 years and above, in both the sexes, in each of the 27 registries, were obtained and presented region-wise.
Cumulative risk computed by NCRP = 100 × (1−exp [−cumulative rate/100]) where, cumulative rate = (5 × ∑ [ASR] × 100)/100,000. The multiplication factor five in the above formula of cumulative rate indicates the 5-year age intervals in ASRs. We computed one in a number of persons likely to develop O&U cancers in the lifetime of 0–75 years in each registry by using a formula 100/cumulative risk. Data on indicators, namely incidence rates, AARs, and one in a number of persons develop cancer, were summarized for both the sexes in each of the cancer registries and presented region-wise in the form of ranges (where applicable).
Joinpoint Regression Program is statistical software developed by the National Cancer Institute for the analysis of trends using joinpoint models, that is, models where several different lines connected together at the “joinpoints.” The software takes trend data (e.g., cancer rates) and fits the simplest joinpoint model that the data allows. The analysis starts with the minimum number of joinpoints (i.e., 0 joinpoint, representing a straight line) and tests whether more joinpoints are statistically significant and added to the model (up to that maximum number). Joinpoint tests of significance use a Monte Carlo permutation method. We used the Joinpoint Regression Programme version 220.127.116.11 to perform joinpoint regression analysis. The joinpoint regression analysis helps in interpreting changing trends of O&U cancers over the years by fitting different time slots by default in its procedure.
| » Results|| |
[Table 1] provides the region-wise incidence rate (AAR) for 0-74 years of age, one in number of persons who develop cancer, and the percentage of O&U cancers. The region-wise analysis revealed the highest susceptibility of the northeastern region with AAR 1.76–29.7 and 1.99–14.68 for men and women, respectively. North region had an AAR of 6.15–13.81 for men and of 4.21–13.09 for women, which was closely followed by a southern region with 5.01–13.76 for men and with 3.13–9.16 for women. Out of all, central regions showed the least susceptibility with AAR of 3.02 for men and 2.78 for women.
|Table 1: Region-wise incidence rate, one out of the total number of persons likely to develop cancer of other and unknown region, out of total cancer cases for men and women in India, during 2012–2014|
Click here to view
Correspondingly, the percentage of O&U cancer out of total cancers was least in the central region with 2.85% for men and 2.46% for women. The percentage of O&U cancers out of total cancers was highest in the northeastern region (23.86%) for both men and women following the northern region (14.23%).
Region-wise analysis for average annual age-specific incidence rates per 100,000 populations, for 2012–2014 for both men and women, revealed that northeast topped in all age groups with men higher than women [Table 2]. Among men, the northeast region had highest incident rates across all age groups (0–14, 15–34, 35–64, and ≥65 years) with (0–6.2), (0–17.7), (22–362.5), and (12–469.3), respectively. The western region recorded the second-highest incident rate (0–4.1) for the age group (0-14 years). Among age groups, 15–34 and 35–64 years, the north region recorded the second-highest incident rates with 4.6–8 and 70.2–156.7, respectively. For the age group 65 years and above, the south region gave the second-highest incident rate statistics (109.4–329.5). The central region recorded the lowest incident rates among all regions for men.
|Table 2: Region-wise average annual age-specific rate per 100,000 populations of other and unknown cancer for men and women in India during 2012–2014|
Click here to view
Age group analysis of women revealed that in the age groups 0–14, 15–34, and 35–64 years, the northeast region recorded highest incident rates 0–3.6, 2.7–12.7, and 25.1–203.9, respectively. South region recorded the highest incident rate for the age group 65 years and above (61.4–205.1) followed by a north region with (56.5–202.1) incident rates. North region recorded the second-highest incident rate for all age groups (0.11).
Data collected showed different trends for different registries and gender. A joinpoint regression analysis of O&U cancers was carried out for six major registries in both men and women. [Table 3] shows the duration considered by default in the joinpoint regression program along with the average annual percentage change (AAPC) for men and women in various cancer registries.
|Table 3: The choice of different years considered in joinpoint regressions of O&U cancers along with annual percentage change (APC) as a trend for different registries|
Click here to view
The AARs of O&U cancers showed positive trends in all PBCRs over the last decade for both men and women except for the female population of Delhi PBCR. For men, Delhi and Barshi were the two registries, which showed an increased incident rate over a decade, but there is a drop in overall AAR. For women, Bangalore and Barshi recorded an overall drop in incident rates despite an increase in the last decade. On the other hand, a unique trend was noticed for the female population of Delhi. They showed a decline in AAR over the last decade.
For men, in Bangalore, the incident rates of O&U cancers had a negative slope from 1986 to 2004 (−0.21) and then an upward trend until 2012 (0.67). In Delhi, a sharp decline was observed in the incident rates from 1988 to 2000 (−0.76) and then a positive trend until 2012 (0.30). Barshi PBCR also registered a negative slope initially for the year 1988–1990 (−5.82) and later registered an upward swing (0.34) for the period of 1990–2012. In Bhopal (0.20), Chennai (0.17), and Mumbai (0.18) upward slope was recorded continuously for the reporting years [Figure 1].
|Figure 1: Joinpoint regression plots on age-adjusted rates (AARs) of cancers of unknown primary origin in six major registries among men in India. (a) Bangalore Cancer registry. (b) Barshi Cancer registry. (c) Bhopal Cancer registry. (d) Delhi Cancer registry. (e) Mumbai Cancer registry. (f) Chennai Cancer registry|
Click here to view
For women, only Bhopal (0.11) and Chennai (0.12) cancer registries showed a positive trend throughout the reporting years. In Bangalore, a decline in AAR was observed from 1986 to 2004 (−0.25) and then an upward trend (0.48) until 2012. Similarly, in Barshi a sharp decline (−5.82) was observed over a time of 1988–1990 which continued to increase until 2012 (0.34). Mumbai registry showed constant AAR (0.02) from the period 1986–1999 but marked a sudden increase (0.35) from 1999–2012. The female population of Delhi had a sharp decline (−4.35) within the span of 2 years (1998–2000) and continued by further decrease but a slow pace (−0.14) [Figure 2].
|Figure 2: Joinpoint regression plots on Age-Adjusted Rates (AARs) of cancers of unknown primary origin in six major registries among women in India. (a) Bangalore Cancer registry. (b) Barshi Cancer registry. (c) Bhopal Cancer registry. (d) Delhi Cancer registry. (e) Mumbai Cancer registry. (f) Chennai Cancer registry. |
Click here to view
| » Discussion|| |
Region-wise analysis of O&U cancers among men and women from 27 PBCRs revealed that the northeastern region had the highest AARs in men and women followed by northern and southern regions. Average ASRs for the period of 2012–2014 revealed that the northeast region topped in all age groups with men higher than women. In joinpoint regression analysis, the incident rates for O&U cancers are in an upward trend for both men and women over the last decade in all PBCRs except for the female population of Delhi PBCR where it is in steady decline. Chennai and Bhopal registries showed a continuous upward trend for both men and women.
Similar to our conclusions many other studies revealed that men are more susceptible than women. The first report of Kolkata PBCR gave some shocking statistics. According to the report, the proportion of O&U cancers in men and women is 13.6% and 10.9%, respectively. The compromised data quality and a certain degree of under-ascertainment are said to be the reason for this huge percentage. Globally, many attempts are made to understand the epidemiology of O&U cancers. One such attempt was done by Muir in 1995, where he studied 26,050 histologically confirmed cases of the unknown primary site. In a study done by Roseman and James on 88 patients of O&U cancers, they found high men to women' ratio, that is, 22:1. Approximately, 3% of the total patient population registered with the Yale-New Haven Hospital Tumour Registry per year, were diagnosed with O&U cancers. A detailed study on the population of Scotland from 1961 to 2010 drew the conclusion that CUP represented 3.9% of all invasive cancers with 50,941 cases during that duration. The number and percentage of CUP registrations had been consistently higher in women, but the percentage decreased in both sexes in the most recent two decades. The same findings were from the previous studies in Southeast Netherlands, South Australia, and Sweden. A study on Norwegian Cancer Registry data (1971–2010) done in 2013 presented that the incidence of O&U cancers increased during the first half of the period followed by a substantial decrease. The United States SEER data in 2013 found that the proportion of all cancers diagnosed as CUP decreased during recent decades and then comprised less than 2% since 2007. Schwartz et al. reported that during 1984–1988 “ill-defined cancer” (ICD-9 195–199) accounted for 7.4% of all cancer deaths. The pattern of an increase in the incidence of CUP followed by a decrease was also observed in South Australia, Sweden,, Finland, Norway,, and to a lesser extent in the United States. Schroten et al. in 2018 represented that in the Netherlands, the incidence of Unknown Primary Tumour (UPT) accounts for 2.5% of all new cancer diagnoses.
Cancer incidence in five continents reported that the average annual AAR (world standard) per 100,000 per year in 1983–1987 for the SEER registries was 9.2 for white men and 6.7 for white women, compared with 11.7 for black men and 8.1 for black women.
The concern regarding the incidence and high mortality related to O&U cancers is not confined to the modern era. Many studies done on the patients of O&U cancers showed low survival rates of patients of O&U cancers. In 1967, Smith et al. reported 53 patients with metastatic CUP site. Of Smiths' patients, 62% died within 6 months of diagnosis, 85% were dead within 2 years, and only 9% lived 5 years or more. Interestingly, their longest survivors had undifferentiated carcinoma. On similar grounds, Keim in 1966 published a study on a series of 19 patients with carcinoma in the neck nodes without a demonstrable primary tumor. Another study in 1970 by Holmes et al. studied 784 registered patients of the Tumour Registry of the University of Kansas Medical Centre between 1944 and 1969. The average age at the diagnosis was 60.2 with the sex ratio of all patients of male : female was 58:42. During the first year after diagnosis, 78.5% of the patients died; an additional 11.1% died during the second year. The survival curve then remains rather flat with 5.1% of the 686 patients living 5 years. Between 5 and 10 years, an additional 1.8% died, leaving 3.3% alive at 10 years.
There are many suggestions and notations made for the changes in the trends or patterns of the incidence rate, and high mortality and low survival rates. Change in the pattern of biology and risk factors for cancer (e.g., tobacco use) or the differences in the use of newer diagnostic techniques or the rapid urbanization post-2000 could be the reason for observed trends. The authors gave several notations for the high mortality and low survival rates such as undiagnosed primary tumor but strong notation was host defense against the tumor. According to the study, there is a possibility that host immunologic defense mechanisms might destroy a small primary tumor but fail to destroy established secondary tumor deposits. The study also mentioned that one of the reasons for the primary tumor not been detected could be its small size.
This study attempts to present the status of O&U cancers in India. Presently, there are 29 PBCRs covering around 11% of the Indian population and as there is no other authentic data source on the magnitude of O&U cancer, PBCRs are the only source.
This report is perhaps a first-ever detailed description of O&U cancer data of ICMR-NCRP since its establishment. O&U cancers have different definitions for various cancer registries. As the primary site was not identified, they are less likely to have a good prognosis due to the lack of suitable treatment. This makes them more important on public health grounds. Apparently, improved diagnostic techniques over the years have not lessened the incidence of this dilemma.
To conclude, though the proportion of O&U cancer is low in India, increasing trends of these cancers are a matter of concern. Advanced and improved diagnostics tools need to introduce along with detail and rigorous clinical research for correct diagnosis.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| » References|| |
World Health Organization. International statistical classification of diseases, injuries and causes of death, 9th
Rev. 1977. Available from: https://apps.who.int/iris/handle/10665/40492
. [Last accessed on 2019 Sept 05].
Abbruzzese JL, Lenzi R, Raber MN, Pathak S, Frost P. The biology of unknown primary tumours. Semin Oncol 1993;20:238-43.
Hillen HF. Unknown primary tumours. Postgrad Med J 2000;76:690-3.
Abbruzzese JL, Abbruzzese MC, Hess KR, Raber MN, Lenzi R, Frost P. Unknown primary carcinoma: Natural history and prognostic factors in 657 consecutive patients. J Clin Oncol 1994;12:1272-80.
Stella G, Senetta R, Cassenti A, Ronco M, Cassoni P. Cancers of unknown primary origin: Current perspectives and future therapeutic strategies. J Transl Med 2012;10;12.
Pavilidis N, Fizazi K. Carcinoma of unknown primary (CUP). Crit Rev Oncol Hematol 2009;69:271-8.
Krishnatreya M, Kataki AC, Sharma JD, Das K. Carcinoma of unknown primary: An early cancer with metastasis for epidemiological purposes? J Med Soc 2013;27:144-6. [Full text]
Fizazi K, Greco FA, Pavlidis N, Pentheroudakis G; ESMO Guidelines Working Group. Cancers of unknown primary site: ESMO clinical practice guidelines for diagnosis, treatment and follow-up. Ann Oncol 2011;22:vi64-8.
World Health Organization. International Statistical Classification of Diseases and Related Health Problems - 10th
Revision. 2016. Available from: https://icd.who.int/browse10/2016/en#/II
. [Last accessed on 2019 Jul 08].
National Cancer Institute. Joinpoint Trend Analysis Software. Surveillance Research Program, Division of Cancer Control and Population Sciences. Available from: https://surveillance.cancer.gov/joinpoint/
. [Last accessed on 2019 Jul 09].
Kim HJ, Fay MP, Feuer EJ, Midthune DN. Permutation tests for joinpoint regression with applications to cancer rates. Stat Med 2000;19:335-51.
Sen U, Sankaranarayanan R, Mandal S, Ramanakumar AV, Parkin DM, Siddiqi M. Cancer patterns in eastern India: The first report of the Kolkata cancer registry. Int J Cancer 2002;100:86-91.
Muir C. Cancer of unknown primary site. Cancer 1995;75:353-6.
Roseman JM, James AG. Metastatic cancers to the neck from undetermined primary sites: Long-term follow-up. J Surg Oncol 1982;19:247-9.
Altman E, Cadman E. An analysis of 1539 patients with cancer of unknown primary site. Cancer 1986;57:120-4.
Brewster DH, Lang J, Bhatti LA, Thomson CS, Oien KA. Descriptive epidemiology of cancer of unknown primary site in Scotland, 1961-2010. Cancer Epidemiol 2014;38:227-34.
van de Wouw AJ, Janssen-Heijnen ML, Coebergh JW, Hillen HF. Epidemiology of unknown primary tumours; incidence and population-based survival of 1285 patients in Southeast Netherlands, 1984-1992. Eur J Cancer 2002;38:409-13.
Luke C, Koczwara B, Karapetis C, Pittman K, Price T, Kotasek D, et al
. Exploring the epidemiological characteristics of cancers of unknown primary site in an Australian population: Implications for research and clinical care. Aust NZ J Public Health 2008;32:383-9.
Shu X, Sundquist K, Sundquist J, Hemminki K. Time trends in incidence, causes of death, and survival of cancer of unknown primary in Sweden. Eur J Cancer Prev 2012;21:281-8.
Brustugun OT, Helland Å. Rapid reduction in the incidence of cancer of unknown primary: A population-based study. Acta Oncol 2014;53:134-7.
Urban D, Rao A, Bressel M, Lawrence YR, Mileshkin L. Cancer of unknown primary: A population-based analysis of temporal change and socioeconomic disparities. Br J Cancer 2013;109:1318-24.
Schwartz E, Kofie VY, Sturgeon SR. Racial differences in ill-defined cancer mortality in the United States and in the District of Columbia. J Epidemiol Community Health 1992;46:390-3.
Randén M, Rutqvist LE, Johansson H. Cancer patients without a known primary: Incidence and survival trends in Sweden 1960-2007. Acta Oncol 2009;48:915-20.
Schroten-Loef C, Verhoeven RHA, de Hingh IHJT, van de Wouw AJ, van Laarhoven HWM, Lemmens VEPP. Unknown primary carcinoma in the Netherlands: Decrease in incidence and survival times remain poor between 2000 and 2012. Eur J Cancer 2018;101:77-86.
Smith PE, Krementz ET, Chapman W. Metastatic cancer without a detectable primary site. Am J Surg 1967;113:633-7.
Keim WF. The occult primary in head and neck surgery. Arch Otolaryngol 1966;84:566-9.
Holmes FF, Fouts TL. Metastatic cancer of unknown primary site. Cancer 1970;26:816-20.
[Figure 1], [Figure 2]
[Table 1], [Table 2], [Table 3]