Deep learning uncertainty quantification for clinical text classification.

INTRODUCTION: Machine learning algorithms are expected to work side-by-side with humans in decision-making pipelines. Thus, the ability of classifiers to make reliable decisions is of paramount importance. Deep neural networks (DNNs) represent the state-of-the-art models to address real-world classification. Although the strength of activation in DNNs is often correlated with the network's confidence, in-depth analyses are needed to establish whether they are well calibrated. METHOD: In this paper, we demonstrate the use of DNN-based classification tools to benefit cancer registries by automating information extraction of disease at diagnosis and at surgery from electronic text pathology reports from the US National Cancer Institute (NCI) Surveillance, Epidemiology, and End Results (SEER) population-based cancer registries. In particular, we introduce multiple methods for selective classification to achieve a target level of accuracy on multiple classification tasks while minimizing the rejection amount-that is, the number of electronic pathology reports for which the model's predictions are unreliable. We evaluate the proposed methods by comparing our approach with the current in-house deep learning-based abstaining classifier. RESULTS: Overall, all the proposed selective classification methods effectively allow for achieving the targeted level of accuracy or higher in a trade-off analysis aimed to minimize the rejection rate. On in-distribution validation and holdout test data, with all the proposed methods, we achieve on all tasks the required target level of accuracy with a lower rejection rate than the deep abstaining classifier (DAC). Interpreting the results for the out-of-distribution test data is more complex; nevertheless, in this case as well, the rejection rate from the best among the proposed methods achieving 97% accuracy or higher is lower than the rejection rate based on the DAC. CONCLUSIONS: We show that although both approaches can flag those samples that should be manually reviewed and labeled by human annotators, the newly proposed methods retain a larger fraction and do so without retraining-thus offering a reduced computational cost compared with the in-house deep learning-based abstaining classifier.

Use of the Decipher genomic classifier among men with prostate cancer in the United States.

BACKGROUND: Management of localized or recurrent prostate cancer since the 1990s has been based on risk stratification using clinicopathological variables, including Gleason score, T stage (based on digital rectal exam), and prostate-specific antigen (PSA). In this study a novel prognostic test, the Decipher Prostate Genomic Classifier (GC), was used to stratify risk of prostate cancer progression in a US national database of men with prostate cancer. METHODS: Records of prostate cancer cases from participating SEER (Surveillance, Epidemiology, and End Results) program registries, diagnosed during the period from 2010 through 2018, were linked to records of testing with the GC prognostic test. Multivariable analysis was used to quantify the association between GC scores or risk groups and use of definitive local therapy after diagnosis in the GC biopsy-tested cohort and postoperative radiotherapy in the GC-tested cohort as well as adverse pathological findings after prostatectomy. RESULTS: A total of 572 545 patients were included in the analysis, of whom 8927 patients underwent GC testing. GC biopsy-tested patients were more likely to undergo active active surveillance or watchful waiting than untested patients (odds ratio [OR] =2.21, 95% confidence interval [CI] = 2.04 to 2.38, P < .001). The highest use of active surveillance or watchful waiting was for patients with a low-risk GC classification (41%) compared with those with an intermediate- (27%) or high-risk (11%) GC classification (P < .001). Among National Comprehensive Cancer Network patients with low and favorable-intermediate risk, higher GC risk class was associated with greater use of local therapy (OR = 4.79, 95% CI = 3.51 to 6.55, P < .001). Within this subset of patients who were subsequently treated with prostatectomy, high GC risk was associated with harboring adverse pathological findings (OR = 2.94, 95% CI = 1.38 to 6.27, P = .005). Use of radiation after prostatectomy was statistically significantly associated with higher GC risk groups (OR = 2.69, 95% CI = 1.89 to 3.84). CONCLUSIONS: There is a strong association between use of the biopsy GC test and likelihood of conservative management. Higher genomic classifier scores are associated with higher rates of adverse pathology at time of surgery and greater use of postoperative radiotherapy.In this study the Decipher Prostate Genomic Classifier (GC) was used to analyze a US national database of men with prostate cancer. Use of the GC was associated with conservative management (ie, active surveillance). Among men who had high-risk GC scores and then had surgery, there was a 3-fold higher chance of having worrisome findings in surgical specimens.

31-Gene Expression Profile Testing in Cutaneous Melanoma and Survival Outcomes in a Population-Based Analysis: A SEER Collaboration.

PURPOSE: The DecisionDx-Melanoma 31-gene expression profile (31-GEP) test is validated to classify cutaneous malignant melanoma (CM) patient risk of recurrence, metastasis, or death as low (class 1A), intermediate (class 1B/2A), or high (class 2B). This study aimed to examine the effect of 31-GEP testing on survival outcomes and confirm the prognostic ability of the 31-GEP at the population level. METHODS: Patients with stage I-III CM with a clinical 31-GEP result between 2016 and 2018 were linked to data from 17 SEER registries (n = 4,687) following registries' operation procedures for linkages. Melanoma-specific survival (MSS) and overall survival (OS) differences by 31-GEP risk category were examined using Kaplan-Meier analysis and the log-rank test. Crude and adjusted hazard ratios (HRs) were calculated using Cox regression model to evaluate variables associated with survival. 31-GEP tested patients were propensity score-matched to a cohort of non-31-GEP tested patients from the SEER database. Robustness of the effect of 31-GEP testing was assessed using resampling. RESULTS: Patients with a 31-GEP class 1A result had higher 3-year MSS and OS than patients with a class 1B/2A or class 2B result (MSS: 99.7% v 97.1% v 89.6%, P < .001; OS: 96.6% v 90.2% v 79.4%, P < .001). A class 2B result was an independent predictor of MSS (HR, 7.00; 95% CI, 2.70 to 18.00) and OS (HR, 2.39; 95% CI, 1.54 to 3.70). 31-GEP testing was associated with a 29% lower MSS mortality (HR, 0.71; 95% CI, 0.53 to 0.94) and 17% lower overall mortality (HR, 0.83; 95% CI, 0.70 to 0.99) relative to untested patients. CONCLUSION: In a population-based, clinically tested melanoma cohort, the 31-GEP stratified patients by their risk of dying from melanoma.

Radon and lung cancer in the pooled uranium miners analysis (PUMA): highly exposed early miners and all miners.

OBJECTIVES: Radon is a ubiquitous occupational and environmental lung carcinogen. We aim to quantify the association between radon progeny and lung cancer mortality in the largest and most up-to-date pooled study of uranium miners. METHODS: The pooled uranium miners analysis combines 7 cohorts of male uranium miners with 7754 lung cancer deaths and 4.3 million person-years of follow-up. Vital status and lung cancer deaths were ascertained between 1946 and 2014. The association between cumulative radon exposure in working level months (WLM) and lung cancer was modelled as the excess relative rate (ERR) per 100 WLM using Poisson regression; variation in the association by temporal and exposure factors was examined. We also examined analyses restricted to miners first hired before 1960 and with <100 WLM cumulative exposure. RESULTS: In a model that allows for variation by attained age, time since exposure and annual exposure rate, the ERR/100 WLM was 4.68 (95% CI 2.88 to 6.96) among miners who were less than 55 years of age and were exposed in the prior 5 to <15 years at annual exposure rates of <0.5 WL. This association decreased with older attained age, longer time since exposure and higher annual exposure rate. In analyses restricted to men first hired before 1960, we observed similar patterns of association but a slightly lower estimate of the ERR/100 WLM. CONCLUSIONS: This new large, pooled study confirms and supports a linear exposure-response relationship between cumulative radon exposure and lung cancer mortality which is jointly modified by temporal and exposure factors.

Cancer statistics for American Indian and Alaska Native individuals, 2022: Including increasing disparities in early onset colorectal cancer.

American Indian and Alaska Native (AIAN) individuals are diverse culturally and geographically but share a high prevalence of chronic illness, largely because of obstacles to high-quality health care. The authors comprehensively examined cancer incidence and mortality among non-Hispanic AIAN individuals, compared with non-Hispanic White individuals for context, using population-based data from the National Cancer Institute, the Centers for Disease Control and Prevention, and the North American Association of Central Cancer Registries. Overall cancer rates among AIAN individuals were 2% higher than among White individuals for incidence (2014 through 2018, confined to Purchased/Referred Care Delivery Area counties to reduce racial misclassification) but 18% higher for mortality (2015 through 2019). However, disparities varied widely by cancer type and geographic region. For example, breast and prostate cancer mortality rates are 8% and 31% higher, respectively, in AIAN individuals than in White individuals despite lower incidence and the availability of early detection tests for these cancers. The burden among AIAN individuals is highest for infection-related cancers (liver, stomach, and cervix), for kidney cancer, and for colorectal cancer among indigenous Alaskans (91.3 vs. 35.5 cases per 100,000 for White Alaskans), who have the highest rates in the world. Steep increases for early onset colorectal cancer, from 18.8 cases per 100,000 Native Alaskans aged 20-49 years during 1998 through 2002 to 34.8 cases per 100,000 during 2014 through 2018, exacerbated this disparity. Death rates for infection-related cancers (liver, stomach, and cervix), as well as kidney cancer, were approximately two-fold higher among AIAN individuals compared with White individuals. These findings highlight the need for more effective strategies to reduce the prevalence of chronic oncogenic infections and improve access to high-quality cancer screening and treatment for AIAN individuals. Mitigating the disparate burden will require expanded financial support of tribal health care as well as increased collaboration and engagement with this marginalized population.

Annual report to the nation on the status of cancer, part 1: National cancer statistics.

BACKGROUND: The American Cancer Society, the Centers for Disease Control and Prevention, the National Cancer Institute, and the North American Association of Central Cancer Registries collaborate to provide annual updates on cancer occurrence and trends in the United States. METHODS: Data on new cancer diagnoses during 2001-2018 were obtained from the North American Association of Central Cancer Registries' Cancer in North America Incidence file, which is comprised of data from Centers for Disease Control and Prevention-funded and National Cancer Institute-funded, population-based cancer registry programs. Data on cancer deaths during 2001-2019 were obtained from the National Center for Health Statistics' National Vital Statistics System. Five-year average incidence and death rates along with trends for all cancers combined and for the leading cancer types are reported by sex, racial/ethnic group, and age. RESULTS: Overall cancer incidence rates were 497 per 100,000 among males (ranging from 306 among Asian/Pacific Islander males to 544 among Black males) and 431 per 100,000 among females (ranging from 309 among Asian/Pacific Islander females to 473 among American Indian/Alaska Native females) during 2014-2018. The trend during the corresponding period was stable among males and increased 0.2% on average per year among females, with differing trends by sex, racial/ethnic group, and cancer type. Among males, incidence rates increased for three cancers (including pancreas and kidney), were stable for seven cancers (including prostate), and decreased for eight (including lung and larynx) of the 18 most common cancers considered in this analysis. Among females, incidence rates increased for seven cancers (including melanoma, liver, and breast), were stable for four cancers (including uterus), and decreased for seven (including thyroid and ovary) of the 18 most common cancers. Overall cancer death rates decreased by 2.3% per year among males and by 1.9% per year among females during 2015-2019, with the sex-specific declining trend reflected in every major racial/ethnic group. During 2015-2019, death rates decreased for 11 of the 19 most common cancers among males and for 14 of the 20 most common cancers among females, with the steepest declines (>4% per year) reported for lung cancer and melanoma. Five-year survival for adenocarcinoma and neuroendocrine pancreatic cancer improved between 2001 and 2018; however, overall incidence (2001-2018) and mortality (2001-2019) continued to increase for this site. Among children (younger than 15 years), recent trends were stable for incidence and decreased for mortality; and among, adolescents and young adults (aged 15-39 years), recent trends increased for incidence and declined for mortality. CONCLUSIONS: Cancer death rates continued to decline overall, for children, and for adolescents and young adults, and treatment advances have led to accelerated declines in death rates for several sites, such as lung and melanoma. The increases in incidence rates for several common cancers in part reflect changes in risk factors, screening test use, and diagnostic practice. Racial/ethnic differences exist in cancer incidence and mortality, highlighting the need to understand and address inequities. Population-based incidence and mortality data inform prevention, early detection, and treatment efforts to help reduce the cancer burden in the United States.

Using ensembles and distillation to optimize the deployment of deep learning models for the classification of electronic cancer pathology reports.

Objective: We aim to reduce overfitting and model overconfidence by distilling the knowledge of an ensemble of deep learning models into a single model for the classification of cancer pathology reports. Materials and Methods: We consider the text classification problem that involves 5 individual tasks. The baseline model consists of a multitask convolutional neural network (MtCNN), and the implemented ensemble (teacher) consists of 1000 MtCNNs. We performed knowledge transfer by training a single model (student) with soft labels derived through the aggregation of ensemble predictions. We evaluate performance based on accuracy and abstention rates by using softmax thresholding. Results: The student model outperforms the baseline MtCNN in terms of abstention rates and accuracy, thereby allowing the model to be used with a larger volume of documents when deployed. The highest boost was observed for subsite and histology, for which the student model classified an additional 1.81% reports for subsite and 3.33% reports for histology. Discussion: Ensemble predictions provide a useful strategy for quantifying the uncertainty inherent in labeled data and thereby enable the construction of soft labels with estimated probabilities for multiple classes for a given document. Training models with the derived soft labels reduce model confidence in difficult-to-classify documents, thereby leading to a reduction in the number of highly confident wrong predictions. Conclusions: Ensemble model distillation is a simple tool to reduce model overconfidence in problems with extreme class imbalance and noisy datasets. These methods can facilitate the deployment of deep learning models in high-risk domains with low computational resources where minimizing inference time is required.

Landscape of mutations in early stage primary cutaneous melanoma: An InterMEL study.

It is unclear why some melanomas aggressively metastasize while others remain indolent. Available studies employing multi-omic profiling of melanomas are based on large primary or metastatic tumors. We examine the genomic landscape of early-stage melanomas diagnosed prior to the modern era of immunological treatments. Untreated cases with Stage II/III cutaneous melanoma were identified from institutions throughout the United States, Australia and Spain. FFPE tumor sections were profiled for mutation, methylation and microRNAs. Preliminary results from mutation profiling and clinical pathologic correlates show the distribution of four driver mutation sub-types: 31% BRAF; 18% NRAS; 21% NF1; 26% Triple Wild Type. BRAF mutant tumors had younger age at diagnosis, more associated nevi, more tumor infiltrating lymphocytes, and fewer thick tumors although at generally more advanced stage. NF1 mutant tumors were frequent on the head/neck in older patients with severe solar elastosis, thicker tumors but in earlier stages. Triple Wild Type tumors were predominantly male, frequently on the leg, with more perineural invasion. Mutations in TERT, TP53, CDKN2A and ARID2 were observed often, with TP53 mutations occurring particularly frequently in the NF1 sub-type. The InterMEL study will provide the most extensive multi-omic profiling of early-stage melanoma to date. Initial results demonstrate a nuanced understanding of the mutational and clinicopathological landscape of these early-stage tumors.

Automatic information extraction from childhood cancer pathology reports.

Objectives: The International Classification of Childhood Cancer (ICCC) facilitates the effective classification of a heterogeneous group of cancers in the important pediatric population. However, there has been no development of machine learning models for the ICCC classification. We developed deep learning-based information extraction models from cancer pathology reports based on the ICD-O-3 coding standard. In this article, we describe extending the models to perform ICCC classification. Materials and Methods: We developed 2 models, ICD-O-3 classification and ICCC recoding (Model 1) and direct ICCC classification (Model 2), and 4 scenarios subject to the training sample size. We evaluated these models with a corpus consisting of 29 206 reports with age at diagnosis between 0 and 19 from 6 state cancer registries. Results: Our findings suggest that the direct ICCC classification (Model 2) is substantially better than reusing the ICD-O-3 classification model (Model 1). Applying the uncertainty quantification mechanism to assess the confidence of the algorithm in assigning a code demonstrated that the model achieved a micro-F1 score of 0.987 while abstaining (not sufficiently confident to assign a code) on only 14.8% of ambiguous pathology reports. Conclusions: Our experimental results suggest that the machine learning-based automatic information extraction from childhood cancer pathology reports in the ICCC is a reliable means of supplementing human annotators at state cancer registries by reading and abstracting the majority of the childhood cancer pathology reports accurately and reliably.

Lung Cancer and Radon: Pooled Analysis of Uranium Miners Hired in 1960 or Later.

BACKGROUND: Despite reductions in exposure for workers and the general public, radon remains a leading cause of lung cancer. Prior studies of underground miners depended heavily upon information on deaths among miners employed in the early years of mine operations when exposures were high and tended to be poorly estimated. OBJECTIVES: To strengthen the basis for radiation protection, we report on the follow-up of workers employed in the later periods of mine operations for whom we have more accurate exposure information and for whom exposures tended to be accrued at intensities that are more comparable to contemporary settings. METHODS: We conducted a pooled analysis of cohort studies of lung cancer mortality among 57,873 male uranium miners in Canada, Czech Republic, France, Germany, and the United States, who were first employed in 1960 or later (thereby excluding miners employed during the periods of highest exposure and focusing on miners who tend to have higher quality assessments of radon progeny exposures). We derived estimates of excess relative rate per 100 working level months (ERR/100 WLM) for mortality from lung cancer. RESULTS: The analysis included 1.9 million person-years of observation and 1,217 deaths due to lung cancer. The relative rate of lung cancer increased in a linear fashion with cumulative exposure to radon progeny (ERR/100 WLM=1.33; 95% CI: 0.89, 1.88). The association was modified by attained age, age at exposure, and annual exposure rate; for attained ages <55 y, the ERR/100 WLM was 8.38 (95% CI: 3.30, 18.99) among miners who were exposed at ≥35 years of age and at annual exposure rates of <0.5 working levels. This association decreased with older attained ages, younger ages at exposure, and higher exposure rates. DISCUSSION: Estimates of association between radon progeny exposure and lung cancer mortality among relatively contemporary miners are coherent with estimates used to inform current protection guidelines. https://doi.org/10.1289/EHP10669.

Optimal vocabulary selection approaches for privacy-preserving deep NLP model training for information extraction and cancer epidemiology.

BACKGROUND: With the use of artificial intelligence and machine learning techniques for biomedical informatics, security and privacy concerns over the data and subject identities have also become an important issue and essential research topic. Without intentional safeguards, machine learning models may find patterns and features to improve task performance that are associated with private personal information. OBJECTIVE: The privacy vulnerability of deep learning models for information extraction from medical textural contents needs to be quantified since the models are exposed to private health information and personally identifiable information. The objective of the study is to quantify the privacy vulnerability of the deep learning models for natural language processing and explore a proper way of securing patients' information to mitigate confidentiality breaches. METHODS: The target model is the multitask convolutional neural network for information extraction from cancer pathology reports, where the data for training the model are from multiple state population-based cancer registries. This study proposes the following schemes to collect vocabularies from the cancer pathology reports; (a) words appearing in multiple registries, and (b) words that have higher mutual information. We performed membership inference attacks on the models in high-performance computing environments. RESULTS: The comparison outcomes suggest that the proposed vocabulary selection methods resulted in lower privacy vulnerability while maintaining the same level of clinical task performance.

A Keyword-Enhanced Approach to Handle Class Imbalance in Clinical Text Classification.

Recent applications ofdeep learning have shown promising results for classifying unstructured text in the healthcare domain. However, the reliability of models in production settings has been hindered by imbalanced data sets in which a small subset of the classes dominate. In the absence of adequate training data, rare classes necessitate additional model constraints for robust performance. Here, we present a strategy for incorporating short sequences of text (i.e. keywords) into training to boost model accuracy on rare classes. In our approach, we assemble a set of keywords, including short phrases, associated with each class. The keywords are then used as additional data during each batch of model training, resulting in a training loss that has contributions from both raw data and keywords. We evaluate our approach on classification of cancer pathology reports, which shows a substantial increase in model performance for rare classes. Furthermore, we analyze the impact of keywords on model output probabilities for bigrams, providing a straightforward method to identify model difficulties for limited training data.

Colorectal Cancer Survivors' Receptivity toward Genomic Testing and Targeted Use of Non-Steroidal Anti-Inflammatory Drugs to Prevent Cancer Recurrence.

Genomic testing and targeted use of non-steroidal anti-inflammatory drugs (NSAIDs) may mitigate cancer recurrence risks. This study examines colorectal cancer (CRC) survivors' interest and receptivity to these strategies. Patients diagnosed with stage I-III CRC in 2004-2012 were recruited through the New Mexico Cancer Registry to complete a cancer survivorship experiences survey. We assessed interest in genomic testing, daily aspirin (ASA) and NSAID use, and receptivity to future daily ASA/NSAIDs. Descriptive statistics and multivariable logistic regression models estimated factors associated with genomic testing interest. Receptivity to future ASA/NSAIDs use was estimated for non-users of ASA/NSAIDs. Among CRC survivors (n = 273), 83% endorsed interest in genomic testing, 25% were ASA users and 47% ASA/NSAIDs users. In our final model, genomic testing interest was associated with being uncoupled [OR = 4.11; 95% CI = 1.49-11.35], low income [OR = 0.35, 95% CI: 0.14-0.88], smoking history [OR = 0.35, 95% CI: 0.14-0.90], low [OR: 0.33, 95% CI: 0.07-1.43] and moderate [OR: 0.26, 95% CI: 0.11-0.61] health literacy, and personal CRC risk worry [OR: 2.86, 95% CI: 1.63-5.02, p = 0.0002]. In our final model, ASA use was associated with age [OR: 1.05, 95% CI: 1.01-1.10] and cardiovascular disease history [OR: 2.42, 95% CI: 1.23-4.73, p = 0.010]. Among non-users ASA/NSAIDs, 83% reported receptivity to ASA/NSAIDs to reduce cancer risks, and no significant correlates were identified. The majority of survivors' expressed genomic testing interest and endorsed receptivity toward ASA/NSAIDs use for cancer risk management. Further research to optimize ASA/NSAIDs use guided by genomic testing is warranted.

Class imbalance in out-of-distribution datasets: Improving the robustness of the TextCNN for the classification of rare cancer types.

In the last decade, the widespread adoption of electronic health record documentation has created huge opportunities for information mining. Natural language processing (NLP) techniques using machine and deep learning are becoming increasingly widespread for information extraction tasks from unstructured clinical notes. Disparities in performance when deploying machine learning models in the real world have recently received considerable attention. In the clinical NLP domain, the robustness of convolutional neural networks (CNNs) for classifying cancer pathology reports under natural distribution shifts remains understudied. In this research, we aim to quantify and improve the performance of the CNN for text classification on out-of-distribution (OOD) datasets resulting from the natural evolution of clinical text in pathology reports. We identified class imbalance due to different prevalence of cancer types as one of the sources of performance drop and analyzed the impact of previous methods for addressing class imbalance when deploying models in real-world domains. Our results show that our novel class-specialized ensemble technique outperforms other methods for the classification of rare cancer types in terms of macro F1 scores. We also found that traditional ensemble methods perform better in top classes, leading to higher micro F1 scores. Based on our findings, we formulate a series of recommendations for other ML practitioners on how to build robust models with extremely imbalanced datasets in biomedical NLP applications.

Antihypertensive medications and risks of recurrence and mortality in luminal, triple-negative, and HER2-overexpressing breast cancer.

PURPOSE: Antihypertensives are commonly prescribed medications and their effect on breast cancer recurrence and mortality is not clear, particularly among specific molecular subtypes of breast cancer: luminal, triple-negative (TN), and HER2-overexpressing (H2E). METHODS: A population-based prospective cohort study of women aged 20-69 diagnosed with a first primary invasive breast cancer between 2004 and 2015 was conducted in the Seattle, Washington and Albuquerque, New Mexico greater metropolitan areas. Multivariable-adjusted Cox proportional hazards regression was used to estimate hazard ratios (HR) and 95% confidence intervals (CI) for risks of breast cancer recurrence, breast cancer-specific mortality, and all-cause mortality associated with hypertension and antihypertensives. RESULTS: In this sample of 2,383 luminal, 1,559 TN, and 615 H2E breast cancer patients, overall median age was 52 (interquartile range, 44-60). Hypertension and current use of antihypertensives were associated with increased risks of all-cause mortality in each subtype. Current use of angiotensin-converting enzyme inhibitors was associated with increased risks of both recurrence and breast cancer-specific mortality among luminal patients (HR: 2.5; 95% CI: 1.5, 4.3 and HR: 1.9; 95% CI: 1.2, 3.0, respectively). Among H2E patients, current use of calcium channel blockers was associated with an increased risk of breast cancer-specific mortality (HR: 1.8; 95% CI: 0.6, 5.4). CONCLUSION: Our findings suggest that some antihypertensive medications may be associated with adverse breast cancer outcomes among women with certain molecular subtypes. Additional studies are needed to confirm these findings.

Clinical follow-up practices after cervical cancer screening by co-testing: A population-based study of adherence to U.S. guideline recommendations.

Failure to follow-up women after abnormal cervical screening could lead to cervical cancers, yet little is known about adherence to recommended follow-up after abnormal co-testing [cytology and high-risk human papillomavirus (hrHPV) testing]. We documented clinical management following cervical screening by co-testing in a diverse population-based setting. A statewide surveillance program for cervical screening, diagnosis, and treatment was used to investigate all cytology, hrHPV and biopsy reports in the state of New Mexico from January 2015 through August 2019. Guideline-adherent follow-up after co-testing required 1) biopsy within 6 months for low-grade cytology if positive for hrHPV, for high-grade cytology irrespective of hrHPV, and for HPV 16/18 positive results irrespective of cytology and; 2) repeat co-testing within 18 months if cytology was negative and hrHPV test was positive (excluding types 16/18). Screening co-tests (2015-2017) for 164,522 women were analyzed using descriptive statistics, Kaplan Meier curves, and pairwise comparisons between groups. Guideline adherence was highest when both cytology and hrHPV tests were abnormal, ranging from 61.7% to 80.3%. Guideline-adherent follow-up was lower for discordant results. Women with high-grade cytology were less likely to receive a timely biopsy when hrHPV-testing was negative (48.1%) versus positive (83.3%) (p < 0.001). Only 47.9% of women received biopsies following detection of HPV16/18 with normal cytology, and 30.8% received no follow-up within 18-months. Among women with hrHPV-positive normal cytology without evidence of HPV 16/18 infection, 51% received no follow-up within 18 months. Provider education and creation of robust recall systems may help ensure appropriate follow-up of abnormal screening results.

A sequential explanatory study of the employment experiences of population-based breast, colorectal, and prostate cancer survivors.

PURPOSE: Cancer treatment often leads to work disruptions including loss of income, resulting in long-term financial instability for cancer survivors and their informal caregivers. METHODS: In this sequential explanatory study, we conducted a cross-sectional survey of employment experiences among ethnically diverse, working-age individuals diagnosed with breast, colorectal, or prostate cancer. Following the survey, we conducted semi-structured interviews with cancer survivors and informal caregivers to explore changes in employment status and coping techniques to manage these changes. RESULTS: Among employed survivors (n = 333), cancer caused numerous work disruptions including issues with physical tasks (53.8%), mental tasks (46.5%) and productivity (76.0%) in the workplace. Prostate cancer survivors reported fewer work disruptions than female breast and male and female colorectal cancer survivors. Paid time off and flexible work schedules were work accommodations reported by 52.6% and 36.3% of survivors, respectively. In an adjusted regression analysis, household income was positively associated with having received a work accommodation. From the qualitative component of the study (survivors n = 17; caregivers n = 11), three key themes emerged: work disruptions, work accommodations, and coping mechanisms to address the disruptions. Survivors and caregivers shared concerns about lack of support at work and resources to navigate issues caused by changes in employment. CONCLUSIONS: This study characterized employment changes among a diverse group of cancer survivors. Work accommodations were identified as a specific unmet need, particularly among low-income cancer survivors. Addressing changes in employment among specific groups of cancer survivors and caregivers is critical to mitigate potential long-term consequences of cancer.