A review of AI and Data Science support for cancer management.

INTRODUCTION: Thanks to improvement of care, cancer has become a chronic condition. But due to the toxicity of treatment, the importance of supporting the quality of life (QoL) of cancer patients increases. Monitoring and managing QoL relies on data collected by the patient in his/her home environment, its integration, and its analysis, which supports personalization of cancer management recommendations. We review the state-of-the-art of computerized systems that employ AI and Data Science methods to monitor the health status and provide support to cancer patients managed at home. OBJECTIVE: Our main objective is to analyze the literature to identify open research challenges that a novel decision support system for cancer patients and clinicians will need to address, point to potential solutions, and provide a list of established best-practices to adopt. METHODS: We designed a review study, in compliance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, analyzing studies retrieved from PubMed related to monitoring cancer patients in their home environments via sensors and self-reporting: what data is collected, what are the techniques used to collect data, semantically integrate it, infer the patient's state from it and deliver coaching/behavior change interventions. RESULTS: Starting from an initial corpus of 819 unique articles, a total of 180 papers were considered in the full-text analysis and 109 were finally included in the review. Our findings are organized and presented in four main sub-topics consisting of data collection, data integration, predictive modeling and patient coaching. CONCLUSION: Development of modern decision support systems for cancer needs to utilize best practices like the use of validated electronic questionnaires for quality-of-life assessment, adoption of appropriate information modeling standards supplemented by terminologies/ontologies, adherence to FAIR data principles, external validation, stratification of patients in subgroups for better predictive modeling, and adoption of formal behavior change theories. Open research challenges include supporting emotional and social dimensions of well-being, including PROs in predictive modeling, and providing better customization of behavioral interventions for the specific population of cancer patients.

A meta-analysis of Hodgkin lymphoma reveals 19p13.3 TCF3 as a novel susceptibility locus.

Recent genome-wide association studies (GWAS) of Hodgkin lymphoma (HL) have identified associations with genetic variation at both HLA and non-HLA loci; however, much of heritable HL susceptibility remains unexplained. Here we perform a meta-analysis of three HL GWAS totaling 1,816 cases and 7,877 controls followed by replication in an independent set of 1,281 cases and 3,218 controls to find novel risk loci. We identify a novel variant at 19p13.3 associated with HL (rs1860661; odds ratio (OR)=0.81, 95% confidence interval (95% CI) = 0.76-0.86, P(combined) = 3.5 × 10(-10)), located in intron 2 of TCF3 (also known as E2A), a regulator of B- and T-cell lineage commitment known to be involved in HL pathogenesis. This meta-analysis also notes associations between previously published loci at 2p16, 5q31, 6p31, 8q24 and 10p14 and HL subtypes. We conclude that our data suggest a link between the 19p13.3 locus, including TCF3, and HL risk.

Hodgkin lymphoma incidence in California Hispanics: influence of nativity and tumor Epstein-Barr virus.

PURPOSE: For classical Hodgkin lymphoma (HL), migrant studies could elucidate contributions of environmental factors (including Epstein-Barr virus (EBV)) to the lower rates in non-whites. Given the well-described etiologic complexity of HL, this research requires a large, immigrant population, such as California Hispanics. METHODS: With 1988-2004 California Cancer Registry data (2,595 Hispanic, 8,637 white HL cases) and tumor cell EBV status on a subset (218 Hispanics, 656 whites), we calculated ethnicity- and nativity-specific HL incidence rates simultaneously by age, sex, and histologic subtype, and tumor cell EBV prevalence. RESULTS: Compared with white rates, Hispanic HL rates were lower overall (70 %) and for nodular sclerosis HL, particularly among young adults (60-65 % for females). However, they were higher among children (200 %) and older adults, and for mixed cellularity HL. Compared with rates in foreign-born Hispanics, rates in US-born Hispanics were higher among young adults (>threefold in females), lower for children and adults over age 70, and consistently intermediate compared with rates in whites. EBV tumor prevalence was 67, 32, and 23 % among foreign-born Hispanics, US-born Hispanics, and whites, respectively, although with variation by age, sex, and histology. CONCLUSIONS: Findings strongly implicate environmental influences, such as nativity-related sociodemographic differences, on HL occurrence. In addition, lower young adult rates and higher EBV prevalence in US-born Hispanics than in whites raise questions about the duration/extent of environmental change for affecting HL rates and also point to ethnic differences in genetic susceptibility. Lesser variation in mixed cellularity HL rates and greater variation in rates for females across groups suggest less modifiable factors interacting with environmental influences.

Cigarette smoking and risk of Hodgkin lymphoma and its subtypes: a pooled analysis from the International Lymphoma Epidemiology Consortium (InterLymph).

BACKGROUND: The etiology of Hodgkin lymphoma (HL) remains incompletely characterized. Studies of the association between smoking and HL have yielded ambiguous results, possibly due to differences between HL subtypes. PATIENTS AND METHODS: Through the InterLymph Consortium, 12 case-control studies regarding cigarette smoking and HL were identified. Pooled analyses on the association between smoking and HL stratified by tumor histology and Epstein-Barr virus (EBV) status were conducted using random effects models adjusted for confounders. Analyses included 3335 HL cases and 14 278 controls. RESULTS: Overall, 54.5% of cases and 57.4% of controls were ever cigarette smokers. Compared with never smokers, ever smokers had an odds ratio (OR) of HL of 1.10 [95% confidence interval (CI) 1.01-1.21]. This increased risk reflected associations with mixed cellularity cHL (OR = 1.60, 95% CI 1.29-1.99) and EBV-positive cHL (OR = 1.81, 95% CI 1.27-2.56) among current smokers, whereas risk of nodular sclerosis (OR = 1.09, 95% CI 0.90-1.32) and EBV-negative HL (OR = 1.02, 95% CI 0.72-1.44) was not increased. CONCLUSION: These results support the notion of etiologic heterogeneity between HL subtypes, highlighting the need for HL stratification in future studies. Even if not relevant to all subtypes, our study emphasizes that cigarette smoking should be added to the few modifiable HL risk factors identified.

Age variation in Hodgkin's disease risk factors in older women: evidence from a population-based case-control study.

Hodgkin's disease (HD), which affects all age groups, has been associated with childhood social class, particularly among adults under age 40. Little is known about social class risk factors in older adults, and the few existing studies have conflicting findings. As part of a population-based case-control study of HD in women, we examined social class risk factors by diagnostic age groups (45-54 years and 55-79 years) corresponding to incidence patterns and by histologic subtypes based on a uniform pathologic review. Among women ages 45-54, cases were more likely to be Catholic, to have lower income and to be taller than controls. Among women ages 55-79, cases tended to have come from small or large childhood households, lived in single-family childhood housing, and had a single rather than shared bedroom at age 11. For the nodular sclerosis (NS) histologic subtype, similar age differences in risk factors were apparent. Comparisons between the NS and non-NS subtypes in women ages 55-79 identified some common risk factors (single-family childhood home, single bedroom at age 11) but others specific to one subtype (childhood household size, adult height for NS; lower maternal education for non-NS). Thus, some social class associations with HD differed between middle-aged and older women, as well as between these groups and younger adults, while others were shared across age groups. Risk also was associated with both higher and lower childhood social class in middle-aged and older women, in contrast with previous findings. None of these patterns was explained entirely by histologic subtype but may reflect age and histology subtype variation in the HD-EBV association.

Age-specific survival after Hodgkin's disease in a population-based cohort (United States).

OBJECTIVE: To examine risk factors for disease-specific survival in young and older adults diagnosed with Hodgkin's disease (HD) in a representative case series of adequate size for detecting effect modification by age group. METHODS: For 5630 young adults (ages 15-44) and 2424 older adults (ages 45 and older) diagnosed with HD and reported to the population-based Surveillance, Epidemiology, and End Results program of the National Cancer Institute between 1983 and 1995, Kaplan-Meier survival curves were constructed and Cox proportional hazards regression used to evaluate the influences of age, sex, race/ethnicity, histologic subtype, Ann Arbor stage at diagnosis, and calendar year on hazard of disease-specific death. RESULTS: The effects of most previously studied risk factors for HD death were different for young and older adults. Age was not associated with disease-specific survival in young adults, but in older adults, 1-year increases in age elevated the relative hazard of HD death by 4 6%. Male sex was related to outcome in young but not older adults, and Ann Arbor stage and B-symptom status exhibited markedly different relationships to survival by age. Older adult patients with and without B-symptoms had different hazards of mortality and had to be assessed separately. CONCLUSIONS: Factors associated with disease-specific survival were different for young and older adults with HD. These findings provide further support for two etiologically and clinically distinct disease entities.

Epidemiologic trends in HIV-associated lymphomas.

Infection with HIV increases the risk of developing non-Hodgkin lymphoma and, to a lesser extent, Hodgkin disease. The introduction of highly active antiretroviral therapy (HAART) in 1996 changed the natural history of HIV disease, but the HIV-infected population also has changed in composition. Accordingly, the epidemiology of HIV-associated lymphomas now differs from that observed in the first 15 years of the HIV epidemic. In populations with access to HAART, reductions in lymphoma risk have been reported for NHL and suggested for Hodgkin disease, but long-term risks are as yet unknown. Lymphomas are increasingly common cancers in persons with HIV and are fatal in most patients, warranting continued attention to their incidence and etiology.

Expert review of the diagnosis and histologic classification of Hodgkin disease in a population-based cancer registry: interobserver reliability and impact on incidence and survival rates.

BACKGROUND: The reliability of Hodgkin disease (HD) diagnosis and histologic classification is an ongoing concern but has not been evaluated in a population-based case series in 20 years. Yet, diagnostic error in cancer registry data used in surveying HD occurrence may produce statistics that misrepresent incidence, mortality, or survival. METHODS: Uniform pathology review was attempted for all 395 women ages 19--79 years with incident HD reported to a population-based cancer registry in 1988--94. Agreement between original registry and review diagnoses was measured with positive predictive values and kappa statistics. Incidence rates and survival probabilities were computed based on registry and review diagnoses. RESULTS: Registry and review diagnosis agreed for 245 of the 362 reviewed cases. Positive predictive values varied by histologic subtype (nodular sclerosis, 95%; lymphocyte predominance, 69%; mixed cellularity, 58%; lymphocyte depletion, 0%; not otherwise specified, 40%), but agreement was good overall (kappa, 0.66, 95% confidence interval, 0.56--0.76). Eleven patients were determined not to have HD; all were older than age 44 years. Hodgkin disease incidence rates differed for original and review diagnoses only in older women, for whom registry rates slightly overestimated incidence. Five-year survival rates did not differ for registry and review data overall or by age group. CONCLUSIONS: For most adult women patients, the diagnosis of HD was confirmed on review, reflecting the very good agreement between registry and review diagnoses for nodular sclerosis, the most common subtype. Thus, cancer registry statistics for this time period can provide accurate estimates of disease patterns for HD overall and for the nodular sclerosis variant. For other histologic subtypes, rates may be unreliable, and HD occurrence overall may be less dependable in populations with larger proportions of these subtypes.

Epstein-Barr virus and survival after Hodgkin disease in a population-based series of women.

BACKGROUND: Epstein-Barr virus (EBV) positive Hodgkin disease (HD), as defined by the presence of EBV genes or gene products in the malignant cells, differs epidemiologically from EBV negative HD. However, survival patterns for EBV-defined HD have not been well studied. To determine if EBV status influenced survival time after HD, the authors investigated a large, population-based series of female patients. METHODS: For 311 female patients living in the Greater San Francisco Bay Area who were aged 19-79 years with HD diagnosed between mid-1988 and 1994, histopathologically rereviewed archived biopsy specimens were assayed for EBV with immunohistochemistry and in situ hybridization. The 53 subjects with EBV positive and the 258 with EBV negative HD were observed for vital status through 1998; overall survival was analyzed with Kaplan-Meier and Cox proportional hazards regression methods. RESULTS: Epstein-Barr virus positive HD patients were older, received diagnosis at a later stage, and were less likely to have nodular sclerosis histology than EBV negative patients. Deaths were reported for 21 (40%) EBV positive and 37 (14%) EBV negative patients. No survival differences were observed between EBV positive and negative women aged 19-44 years, but survival was significantly poorer in women aged 45-79 years with EBV positive HD. Regression analysis confirmed this strong negative effect of EBV positive status on survival (hazard ratio for death, 3.0; 95% confidence interval, 1.5-6.2) as unrelated to age, stage at diagnosis, or tumor histology. CONCLUSIONS: This study found a marked survival disadvantage for EBV positive HD in older but not young adult women. These findings suggest influences of both EBV status and age on HD survival, as well as pathogenesis.

Epstein-barr virus-associated malignancies: epidemiologic patterns and etiologic implications.

Epstein-Barr virus (EBV), a ubiquitous B-lymphotrophic herpesvirus, has been found in the tumor cells of a heterogeneous group of malignancies (Burkitt's lymphoma, lymphomas associated with immunosuppression, other non-Hodgkin's lymphomas, Hodgkin's disease, nasopharyngeal carcinoma, gastric adenocarcinoma, lymphoepithelioma-like carcinomas, and immunodeficiency-related leiomyosarcoma). As the epidemiologic characteristics of these cancers have not been considered together, this review seeks to relate their incidence patterns and risk factors to EBV biology and virus-host interaction in an attempt to help elucidate factors involved in EBV-related carcinogenesis. We include a brief review of EBV virology and primary infection to provide a biologic context for considering the epidemiology, summarize the most salient epidemiologic features of each malignancy, synthesize epidemiologic data by risk factor to uncover commonalities and informative contrasts across the diseases, and propose hypotheses regarding etiologic mechanisms, based on the possible effect of the risk factors at various stages in the viral life cycle.

Comparison of methods for classifying Hispanic ethnicity in a population-based cancer registry.

The accuracy of ethnic classification can substantially affect ethnic-specific cancer statistics. In the Greater Bay Area Cancer Registry, which is part of the Surveillance, Epidemiology, and End Results (SEER) Program and of the statewide California Cancer Registry, Hispanic ethnicity is determined by medical record review and by matching to surname lists. This study compared these classification methods with self-report. Ethnic self-identification was obtained by surveying 1,154 area residents aged 20-89 years who were diagnosed with cancer in 1990 and were reported to the registry as being Hispanic or White non-Hispanic. Predictive value positive, sensitivity, and relative bias were used to assess the accuracy of Hispanic classification by medical record and surname. Among those persons classified as Hispanic by either or both of these sources, only two-thirds agreed (predictive value positive = 66%), and many self-identified Hispanics were classified incorrectly (sensitivity = 68%). Classification based on either medical record or surname alone had a lower sensitivity (59% and 61%, respectively) but a higher predictive value positive (77% and 70%, respectively). Ethnic classification by medical record alone resulted in an underestimate of Hispanic cancer cases and incidence rates. Bias was reduced when medical records and surnames were used together to classify cancer cases as Hispanic.

Accuracy of racial classification of Vietnamese patients in a population-based cancer registry.

Racial classification of Asian subgroups is increasingly important for health statistics, given the growing Asian-American populations. This study reports the reliability of racial classification of Vietnamese in population-based cancer registry data from northern California. From the Greater Bay Area Cancer Registry, we selected 2240 persons diagnosed with cancer in 1989-1992 and whom the registry considered Vietnamese by birthplace and/or registry race and/or surname, or who were Southeast Asian or Chinese by race. One thousand ninety persons (49%) were interviewed. Sensitivity and predictive value positive, and cancer incidence rates, were calculated using different combinations of the classification factors (birthplace, registry race, and name). By registry-reported race alone, 74% of those the registry classified as Vietnamese agreed with this classification on interview, while 90% of those identifying themselves as Vietnamese were so classified. With classification based on 2 of 3 factors, 78% of those classified as Vietnamese agreed, and 91% of self-reported Vietnamese were correctly classified. Misclassification was associated with age, sex, year of immigration, education, and language use. Registry-based annual age-adjusted all-site cancer incidence rates per 100,000 for Vietnamese were 287.7 for males and 221.3 for females. Rates adjusted for self-reported ethnicity were 242.8 (male) and 213.7 (female). Registry classification of Vietnamese is currently problematic. Approximately 20% of cancer cases classified as Vietnamese are probably not Vietnamese. The higher incidence rates for Vietnamese in the United States than in Vietnam partly may reflect such classification error.

Adjustment of cancer incidence rates for ethnic misclassification.

Although ethnic population counts measured by the United States Census are based on self-identification, the same is not necessarily true of cases reported to cancer registries. The use of different ethnic classification methods for numerators and denominators may therefore lead to biased estimates of cancer incidence rates. The extent of such misclassification may be assessed by conducting an ethnicity survey of cancer patients and estimating the proportion misclassified using double sampling models that account for sample stratification. For two ethnic categories, logistic regression may be used to model self-identified ethnicity as a function of demographic variables and the fallible classification method. Incidence rates then may be adjusted for misclassification using regression results to estimate the number of cancer cases of a given age, sex, and site in each self-identified ethnic group. An example is given using this method to estimate ethnic misclassification of San Francisco Bay area Hispanic cancer patients diagnosed in 1990. Results suggest that the number of cancer cases reported as Hispanic is an underestimate of the number of cases self-identified as Hispanic, resulting in an underestimate of Hispanic cancer rates.

Absence of Epstein-Barr virus EBER-1 transcripts in an epidemiologically diverse group of breast cancers.

Epstein-Barr virus (EBV), a ubiquitous herpesvirus associated with certain lymphomas and carcinomas, has been identified within the malignant cells of a small proportion of breast tumors. As breast cancer is a very common malignancy in women, a pathogenetic role of EBV for even a subgroup of patients could have important implications for etiology and prevention. Therefore, we attempted to confirm the EBV-breast cancer association by exploring it in a representative case series stratified by characteristics that modify breast cancer risk. We studied a sample of 97 female and 28 male patients identified from a US population-based cancer registry. Patients were selected randomly within age, sex, ethnicity and tumor estrogen-receptor status groups. With their archived tumor tissues, we examined EBV presence using in situ hybridization for the EBER-1 transcript. In the 107 technically adequate specimens, we did not detect this viral transcript in any tumors, including one from a woman who also had an EBER-positive nasopharyngeal carcinoma. Our uniformly negative findings are extremely unlikely to have occurred by chance and cannot be attributed to selective sampling, as our study group included persons at diverse risk for breast cancer. We conclude that the EBV EBER-1 transcript is not commonly expressed in breast cancer, based on a broadly representative case series, though we cannot exclude an association of EBV within a particular population subgroup.

Predictors of misclassification of Hispanic ethnicity in a population-based cancer registry.

PURPOSE: Hispanic ethnicity is often used as a category for calculating population-based rates or assessing risk of epidemiologic studies. However, ethnic misclassification can lead to false conclusions unless the extent of misclassification and the characteristics of those misclassified are understood. METHODS: This study explored determinants of ethnic misclassification in a sample of 1154 cancer cases in the San Francisco-Oakland cancer registry, where ethnic classification is based on surname or medical record report. We compared the following: correctly classified Hispanics, persons classified as Hispanic who self-identified as non-Hispanic, and persons classified as non-Hispanic who self-identified as Hispanic. RESULTS: Among men classified as Hispanic, those most likely to self-identify as non-Hispanic did not speak Spanish, had non-Spanish surnames, and were recent immigrants. Women misclassified as Hispanic did not speak Spanish or have Spanish maiden names, nor did they have mothers with Spanish maiden names. Persons who called themselves Hispanic, but were misclassified by the registry, were likely to be non-Spanish speaking college-education males. CONCLUSIONS: Researchers using ethnicity should be aware of how ethnicity was determined and how this classification may bias or confound their results.

Epstein-Barr virus-associated Hodgkin's disease: epidemiologic characteristics in international data.

Hodgkin's disease (HD) has long been suspected to have an infectious precursor, and indirect evidence has implicated Epstein-Barr virus (EBV), a ubiquitous herpesvirus, as a causal agent. Recent molecular studies using EBER in situ hybridization or latency membrane protein-I (LMP-I) immunohistochemistry have identified EBV latent infection in up to 50% of HD tumors. However, the epidemiologic features of these cases have not been examined in detail. To explore the epidemiology of EBV-positive HD so as to understand the role of EBV in HD etiology more clearly, this project accumulated patient data from 14 studies that had applied these EBV assays to HD tumors. With information on age at diagnosis, sex, ethnicity, histologic subtype, country of residence, clinical stage and EBV tumor status from 1,546 HD patients, we examined risk for EBV-positive disease using logistic regression. Forty percent of subjects had EBV-positive tumors, and EBV prevalence varied significantly across groups defined by the study variables. Odds ratios (OR) for EBV-associated HD were significantly elevated for Hispanics vs. whites (OR = 4.1), mixed cellularity vs. nodular sclerosis histologic subtypes (OR = 7.3, 13.4, 4.9 for ages 0-14, 15-49, 50+ years), children from economically less-developed vs. more-developed regions and young adult males vs. females (OR = 2.5). These findings suggest that age, sex, ethnicity and the physiologic effects of poverty may represent biologic modifiers of the EBV association and confirm that this association is strongly but variably linked to histologic subtype. The data augment biologic evidence that EBV is actively involved in HD pathogenesis in some cases but describe epidemiologic complexity in this process.

The epidemiology of Hodgkin's disease.

Much of the epidemiological heterogeneity of HD incidence reflects the behaviour of the NS subtype, at least in the USA. Incidence variation across races (except Asians) and time periods is most marked in this subtype. In young adults with HD, there is compelling evidence for social class modification of risk consistent with an infectious aetiology; limited data suggest that this effect occurs within the NS subtype, but considerable evidence indicates that it does not primarily involve EBV infection. Findings from familial aggregation studies and HLA associations point to inherited susceptibility to this subtype. Despite little sex difference for NS in young adulthood in the latest incidence data, parity nevertheless appears to be protective against this subtype for women. Therefore, the greater increase for females than males in the incidence of young-adult NS in recent years may reflect the impact of population trends towards later childbearing and lower parity. This change, as well as the concomitant smaller family sizes and growing affluence, could explain part of the burgeoning incidence of NS in young adults in the USA. These observations suggest that NS in young adults constitutes a separate disease, probably of infectious origin. The incongruous occurrence of this subtype in older adults, and the presence of EBV in some NS cases, could reflect heterogeneity within NS, for example, representing features of the cellular phase of NS (Cozen et al, 1992; Medeiros and Greiner, 1995). For the non-NS subtypes, many of the factors that predict risk of NS may also be relevant. Patterns of social class determinants in children and older adults, the age groups at risk for MC, support involvement of an infectious precursor given intense exposure, and EBV is a likely candidate, based on its high prevalence in these groups. However, little aetiological research has been directed explicitly at the non-NS subtypes. Considerable effort has gone into exploring an infectious aetiology of HD. Recently, this line of investigation has moved beyond social class determinants to molecular epidemiological studies of EBV and, to a lesser degree, other potentially involved viruses. The roles of genetic susceptibility and sex hormones also represent promising areas for exploration, particularly in their possible interaction with infectious agents and other environmental factors. Ultimately, clearer epidemiological understanding of HD will be aided by more precise classification of this disease at the molecular level.