TCGA-Reports: A machine-readable pathology report resource for benchmarking text-based AI models

In cancer research, pathology report text is a largely untapped data source. Pathology reports are routinely generated, more nuanced than structured data, and contain added insight from pathologists. However, there are no publicly available datasets for benchmarking report-based models. Two recent advances suggest the urgent need for a benchmark dataset. First, improved optical character recognition (OCR) techniques will make it possible to access older pathology reports in an automated way, increasing the data available for analysis. Second, recent improvements in natural language processing (NLP) techniques using artificial intelligence (AI) allow more accurate prediction of clinical targets from text. We apply state-of-the-art OCR and customized post-processing to report PDFs from The Cancer Genome Atlas, generating a machine-readable corpus of 9,523 reports. Finally, we perform a proof-of-principle cancer-type classification across 32 tissues, achieving 0.992 average AU-ROC. This dataset will be useful to researchers across specialties, including research clinicians, clinical trial investigators, and clinical NLP researchers.

Benchmark Pathology Report Text Corpus with Cancer Type Classification

In cancer research, pathology report text is a largely un-tapped data source. Pathology reports are routinely generated, more nuanced than structured data, and contain added insight from pathologists. However, there are no publicly-available datasets for benchmarking report-based models. Two recent advances suggest the urgent need for a benchmark dataset. First, improved optical character recognition (OCR) techniques will make it possible to access older pathology reports in an automated way, increasing data available for analysis. Second, recent improvements in natural language processing (NLP) techniques using AI allow more accurate prediction of clinical targets from text. We apply state-of-the-art OCR and customized post-processing to publicly available report PDFs from The Cancer Genome Atlas, generating a machine-readable corpus of 9,523 reports. We perform a proof-of-principle cancer-type classification across 32 tissues, achieving 0.992 average AU-ROC. This dataset will be useful to researchers across specialties, including research clinicians, clinical trial investigators, and clinical NLP researchers.

Function in Cancer Patients: Disease and Clinical Determinants

Patients with cancer often experience changes in function during and after treatment but it is not clear what cancer types, and associated clinical factors, affect function. This study evaluated patient-reported functional impairments between specific cancer types and risk factors related to disease status and non-cancer factors. A cross-sectional study evaluating 332 individuals referred to cancer rehabilitation clinics was performed at six U.S. hospitals. The PROMIS Cancer Function Brief 3D Profile was used to assess functional outcomes across the domains of physical function, fatigue, and social participation. Multivariable modeling showed an interaction between cancer type and cancer status on the physical function and social participation scales. Subset analyses in the active cancer group showed an effect by cancer type for physical function (p < 0.001) and social participation (p = 0.008), but no effect was found within the non-active cancer subset analyses. Brain, sarcoma, prostate, and lymphoma were the cancers associated with lower function when disease was active. Premorbid neurologic or musculoskeletal impairments were found to be predictors of lower physical function and social participation in those with non-active cancer; cancer type did not predict low function in patients with no evidence of disease. There was no differential effect of cancer type on fatigue, but increased fatigue was significantly associated with lower age (0.027), increased body mass index (p < 0.001), premorbid musculoskeletal impairment (p < 0.015), and active cancer status (p < 0.001). Anticipatory guidance and education on the common impairments observed with specific cancer types and during specific stages of cancer care may help improve/support patients and their caregivers as they receive impairment-driven cancer rehabilitation care.

Exploring engagement in cervical cancer prevention services among Haitian women in Haiti and in the United States: a scoping review

PROBLEM IDENTIFICATION

Haitian women in Haiti and in the United States experience a disproportionate burden of cervical cancer, however their uptake of cervical cancer prevention services remains concerningly low.

LITERATURE SEARCH

A comprehensive search on bibliographic databases coupled with a grey literature search was conducted. A total of 401 studies were identified, with 28 studies retained after following Arksey and O'Malley's Scoping Review Guidelines.

DATA EVALUATION/SYNTHESIS

Knowledge levels of HPV and cervical cancer, along with preventative measures was alarmingly low. Traditional health practices, cultural worldviews, and social networks had an influence on the uptake of cervical cancer prevention. Health systems barriers were found to be a prevalent barrier among Haitian women in the U.S.

CONCLUSIONS

Future health promotion interventions developed for Haitian women must address personal, cultural, social, and structural factors with an emphasis on modifying knowledge and beliefs to improve engagement in cervical cancer prevention behaviors.

Family caregiving experiences with hospice lung cancer patients compared to other cancer types

BACKGROUND

Family caregivers of cancer patients are very involved in communication with healthcare teams; however, little is known about their experiences. Limited information is known about how the type of cancer patients have impact caregiving experiences.

OBJECTIVES

This study seeks to compare the caregiving experience of caregivers of hospice lung cancer patients with hospice caregivers of patients with all other cancer types.

METHOD

This study is based on a secondary analysis of data generated from a parent study evaluating a behavioral intervention with caregivers of hospice cancer patients.

RESULTS

When comparing caregiving experiences by patient diagnosis, significant differences were found in caregivers of hospice lung cancer demographics and experiences with caregiver-centered communication. Specifically, caregivers of lung cancer patients have significantly more trouble with exchange of information, fostering relationships, and decision making with their hospice team.

CONCLUSION

More research is needed to understand the impact of lung cancer on caregiver centered communication and the necessary interventions required to address these issues.

Cancer Profiles, Times to Treatment, and Survival for Adolescents and Young Adults: Comparisons with Children and Older Adults in New South Wales, Australia

Purpose: To compare cancer types, stages, times to treatment, and survival for adolescent and young adults (AYAs) 15-24 years of age with other cancer patients <40 years. Methods: New South Wales Cancer Registry and treatment data were linked to explore differences in cancer type, stage, time to treatment, and survival between AYAs, children, and adults. Multivariable logistic regression and competing-risk regression were adjusted for sociodemographic, diagnostic period, and clinical characteristics. Results: Most common cancers in AYAs and adults were carcinomas compared with leukemias in children. Advanced (regional and distant) stage applied to 33% of AYA solid cancers, which was similar to adult stages, but lower than the 40% for children (adjusted odds ratio 1.21, 95% confidence interval [CI] 1.01-1.47). Proportions starting treatment ≤60 days from diagnosis were 93% for AYAs and children, and 94% for adults, with higher adjusted odds of starting ≤60 days of 1.39 (95% CI 1.11-1.73) for children and 1.23 (95% CI 1.06-1.44) for adults. Five-year disease-specific survival was 90% for AYAs and adults, and 87% for children. The adjusted subhazard ratio for children compared with AYAs was 0.67 (95% CI 0.52-0.88). Age differences in cancer stage, treatment start, and cancer survival varied by cancer type. Conclusions: After adjusting for cancer type, diagnostic period, and sociodemographic characteristics, AYAs had less advanced solid tumors than children; fewer AYAs were treated within 60 days than children and adults; and AYA survival was lower than for children. The potential for residual confounding from leukemia type and other confounders needs further analysis with larger Australia-wide cohorts.

Prediction of Chemosensitivity in Multiple Primary Cancer Patients Using Machine Learning

BACKGROUND/AIM

Many cancer patients face multiple primary cancers. It is challenging to find an anticancer therapy that covers both cancer types in such patients. In personalized medicine, drug response is predicted using genomic information, which makes it possible to choose the most effective therapy for these cancer patients. The aim of this study was to identify chemosensitive gene sets and compare the predictive accuracy of response of cancer cell lines to drug treatment, based on both the genomic features of cell lines and cancer types.

MATERIALS AND METHODS

In this study, we identified a gene set that is sensitive to a specific therapeutic drug, and compared the performance of several predictive models using the identified genes and cancer types through machine learning (ML). To this end, publicly available gene expression datasets and drug sensitivity datasets of gastric and pancreatic cancers were used. Five ML algorithms, including linear discriminant analysis, classification and regression tree, k-nearest neighbors, support vector machine and random forest, were implemented.

RESULTS

The predictive accuracy of the cancer type models were 0.729 to 0.763 on the training dataset and 0.731 to 0.765 on the testing dataset. The predictive accuracy of the genomic prediction models was 0.818 to 1.0 on the training dataset and 0.759 to 0.896 on the testing dataset.

CONCLUSION

Performance of the specific gene models was much better than those of the cancer type models using the ML methods. Therofore, the most effective therapeutic drug can be chosen based on the expression of specific genes in patients with multiple primary cancers, regardless of cancer types.

Identifying Robust Microbiota Signatures and Interpretable Rules to Distinguish Cancer Subtypes

Cancer can be generally defined as a cluster of systematic diseases triggered by abnormal cell proliferation and growth. With the development of biological sciences and biotechnologies, the etiology of cancer is partially revealed, including some of the most substantial pathogenic factors [either endogenous (genetics) or exogenous (environmental)]. However, some remaining factors that contribute to the tumorigenesis but have not been analyzed and discussed in detail remain. For instance, some typical correlations between microorganisms and tumorigenesis have been reported already, but previous studies are just sporadic studies on single microorganism–cancer subtype pairs and do not explain and validate the specific contribution of microbiome on tumorigenesis. On the basis of the systematic microbiome analyses of blood and cancer-associated tissues in cancer patients/controls in public domain, we performed interpretable analyses. We identified several core regulatory microorganisms that contribute to the classification of multiple tumor subtypes and established quantitative predictive models for interpretable prediction by using multiple machine learning methods. We also compared the optimal features (microorganisms) and rules identified from microbiome profiles processed using the Kraken and the SHOGUN. Collectively, our study identified new microbiome signatures and their interpretable classification rules for cancer discrimination and carried out reliable methodological comparison for robust cancer microbiome analyses, thereby promoting the development of tumor etiology at the microbiome level.

Analysis of Expression Pattern of snoRNAs in Different Cancer Types with Machine Learning Algorithms

Small nucleolar RNAs (snoRNAs) are a new type of functional small RNAs involved in the chemical modifications of rRNAs, tRNAs, and small nuclear RNAs. It is reported that they play important roles in tumorigenesis via various regulatory modes. snoRNAs can both participate in the regulation of methylation and pseudouridylation and regulate the expression pattern of their host genes. This research investigated the expression pattern of snoRNAs in eight major cancer types in TCGA via several machine learning algorithms. The expression levels of snoRNAs were first analyzed by a powerful feature selection method, Monte Carlo feature selection (MCFS). A feature list and some informative features were accessed. Then, the incremental feature selection (IFS) was applied to the feature list to extract optimal features/snoRNAs, which can make the support vector machine (SVM) yield best performance. The discriminative snoRNAs included HBII-52-14, HBII-336, SNORD123, HBII-85-29, HBII-420, U3, HBI-43, SNORD116, SNORA73B, SCARNA4, HBII-85-20, etc., on which the SVM can provide a Matthew’s correlation coefficient (MCC) of 0.881 for predicting these eight cancer types. On the other hand, the informative features were fed into the Johnson reducer and repeated incremental pruning to produce error reduction (RIPPER) algorithms to generate classification rules, which can clearly show different snoRNAs expression patterns in different cancer types. The analysis results indicated that extracted discriminative snoRNAs can be important for identifying cancer samples in different types and the expression pattern of snoRNAs in different cancer types can be partly uncovered by quantitative recognition rules.

Plasma Tumor M2-Pyruvate Kinase Levels in Different Cancer Types

BACKGROUND/AIM

Tumor M2-pyruvate kinase (M2-PK) is up-regulated in proliferating tissues. It has been shown that tumor M2-PK is detectable and quantifiable in the stool and plasma of patients with colorectal cancer (CRC). Tumor M2-PK has been extensively studied in gastrointestinal tumors but its role in other cancer types has not yet been deeply evaluated. The aim of the study was to determine and compare plasma tumor M2-PK levels in different cancer types.

MATERIALS AND METHODS

All patients undergoing diagnostics for cancer at our Hospital during 2011 were included in the study (n=139). Plasma tumor M2-PK concentration was analyzed by an enzyme-linked immunosorbent assay.

RESULTS

The different cancer types found in the study were: 60 colorectal, 43 breast, 8 lung, 5 prostatic, 4 ovarian and the remaining 19 cases were other uncommon tumor types. Most tumors had high concentrations of tumor M2-PK; prostatic, pharyngeal and testicular tumors had levels lower than or near the cut-off. Plasma tumor M2-PK levels were significantly higher in patients with distant metastases and stage IV by TNM.

CONCLUSION

Plasma tumor M2-PK is not a specific marker for CRC and is elevated in many other types of cancers, including breast, lung, ovarian, and thyroid. Small amounts are found in prostatic, pharyngeal and testicular tumors.

Do cancer and treatment type affect distress?

OBJECTIVE

We examined differences in distress levels and Distress Thermometer (DT) cutoff scores between different cancer types. The effect of socio-demographic and illness-related variables on distress was also examined.

METHODS

One thousand three hundred fifty patients (response = 51%) completed questions on socio-demographic and illness-related variables, the Dutch version of the DT and Problem List, and the Hospital Anxiety and Depression Scale. Receiver operating characteristics analyses were performed to determine cancer specific cutoff scores. Univariate and multivariate effects of socio-demographic and illness-related variables (including cancer type) on distress were examined.

RESULTS

Prostate cancer patients reported significantly lower DT scores (M = 2.5 ± 2.5) and the cutoff score was lower (≥ 4) than in patients with most other cancer types (M varied between 3.4 and 5.1; cutoff ≥ 5). Multivariate analyses (F = 10.86, p < .001, R(2) = 0.08) showed an independent significant effect of four variables on distress: intensive treatment (β = .10, any (combination of) treatment but surgery only and 'wait and see'); a non-prostate cancer type (β = -.17); the interaction between gender and age (β = -.12, highest distress in younger women as compared with older women and younger and older men); and the interaction between cancer type and treatment intensity (β = .08, lowest scores in prostate cancer patients receiving non-intensive treatment as compared with their counterparts).

CONCLUSIONS

Distress and cutoff score in prostate cancer patients were lower than in patients with other cancer types. Additionally, younger women and patients receiving treatment other than surgery only or 'wait and see' are at risk for higher distress. These results can help identify patients possibly in need of referral to professional psychosocial and/or allied health care.