Predicting cancer prognosis using interactive online tools: a systematic review and implications for cancer care providers

Impact of an artificial intelligence based model to predict non-transplantable recurrence among patients with hepatocellular carcinoma

OBJECTIVE

We sought to develop Artificial Intelligence (AI) based models to predict non-transplantable recurrence (NTR) of hepatocellular carcinoma (HCC) following hepatic resection (HR).

METHODS

HCC patients who underwent HR between 2000-2020 were identified from a multi-institutional database. NTR was defined as recurrence beyond Milan Criteria. Different machine learning (ML) and deep learning (DL) techniques were used to develop and validate two prediction models for NTR, one using only preoperative factors and a second using both preoperative and postoperative factors.

RESULTS

Overall, 1763 HCC patients were included. Among 877 patients with recurrence, 364 (41.5%) patients developed NTR. An ensemble AI model demonstrated the highest area under ROC curves (AUC) of 0.751 (95% CI: 0.719-0.782) and 0.717 (95% CI:0.653-0.782) in the training and testing cohorts, respectively which improved to 0.858 (95% CI: 0.835-0.884) and 0.764 (95% CI: 0.704-0.826), respectively after incorporation of postoperative pathologic factors. Radiologic tumor burden score and pathological microvascular invasion were the most important preoperative and postoperative factors, respectively to predict NTR. Patients predicted to develop NTR had overall 1- and 5-year survival of 75.6% and 28.2%, versus 93.4% and 55.9%, respectively, among patients predicted to not develop NTR (p < 0.0001).

CONCLUSION

The AI preoperative model may help inform decision of HR versus LT for HCC, while the combined AI model can frame individualized postoperative care (https://altaf-pawlik-hcc-ntr-calculator.streamlit.app/).

Modified integrated tumor burden, liver function, systemic inflammation, and tumor biology score to predict long-term outcomes after resection for hepatocellular carcinoma

BACKGROUND

A preoperative predictive score for hepatocellular carcinoma (HCC) can help stratify patients who undergo resection relative to long-term outcomes and tailor treatment strategies.

METHODS

Patients who underwent curative-intent hepatectomy for HCC between 2000 and 2020 were identified from an international multi-institutional database. A risk score (mFIBA) was developed using an Eastern cohort and then validated using a Western cohort.

RESULTS

Among 957 patients, 443 and 514 patients were included from the Eastern and Western cohorts, respectively. On multivariable analysis, alpha-feto protein (HR1.97, 95%CI 1.42-2.72), neutrophil-to-lymphocyte ratio (HR1.74, 95%CI 1.28-2.38), albumin-bilirubin grade (HR1.66, 95%CI 1.21-2.28), and imaging tumor burden score (HR1.25, 95%CI 1.12-1.40) were associated with OS. The c-index in the Eastern test and Western validation cohorts were 0.69 and 0.67, respectively. Notably, mFIBA score outperformed previous HCC staging systems. 5-year OS incrementally decreased with an increase in mFIBA. On multivariable Cox regression analysis, the mFIBA score was associated with worse OS (HR1.18, 95%CI 1.13-1.23) and higher risk of recurrence (HR1.16, 95%CI 1.11-1.20). An easy-to-use calculator of the mFIBA score was made available online (https://yutaka-endo.shinyapps.io/mFIBA_score/).

DISCUSSION

The online mFIBA calculator may help surgeons with clinical decision-making to individualize perioperative treatment strategies for patients undergoing resection of HCC.

Development and validation of risk of CPS decline (RCD): a new prediction tool for worsening cognitive performance among home care clients in Canada

BACKGROUND

To develop and validate a prediction tool, or nomogram, for the risk of a decline in cognitive performance based on the interRAI Cognitive Performance Scale (CPS).

METHODS

Retrospective, population-based, cohort study using Canadian Resident Assessment Instrument for Home Care (RAI-HC) data, collected between 2010 and 2018. Eligible home care clients, aged 18+, with at least two assessments were selected randomly for model derivation (75%) and validation (25%). All clients had a CPS score of zero (intact) or one (borderline intact) on intake into the home care program, out of a possible score of six. All individuals had to remain as home care recipients for the six months observation window in order to be included in the analysis. The primary outcome was any degree of worsening (i.e., increase) on the CPS score within six months. Using the derivation cohort, we developed a multivariable logistic regression model to predict the risk of a deterioration in the CPS score. Model performance was assessed on the validation cohort using discrimination and calibration plots.

RESULTS

We identified 39,292 eligible home care clients, with a median age of 79.0 years, 62.3% were female, 38.8% were married and 38.6% lived alone. On average, 30.3% experienced a worsening on the CPS score within the six-month window (i.e., a change from 0 or 1 to 2, 3, 4, 5, or 6). The final model had good discrimination (c-statistic of 0.65), with excellent calibration.

CONCLUSIONS

The model accurately predicted the risk of deterioration on the CPS score over six months among home care clients. This type of predictive model may provide useful information to support decisions for home care clinicians who use interRAI data internationally.

A New Personalized Oral Cancer Survival Calculator to Estimate Risk of Death From Both Oral Cancer and Other Causes

Importance

Standard cancer prognosis models typically do not include much specificity in characterizing competing illnesses or general health status when providing prognosis estimates, limiting their utility for individuals, who must consider their cancer in the context of their overall health. This is especially true for patients with oral cancer, who frequently have competing illnesses.

Objective

To describe a statistical framework and accompanying new publicly available calculator that provides personalized estimates of the probability of a patient surviving or dying from cancer or other causes, using oral cancer as the first data set.

Design, Setting, and Participants

The models used data from the Surveillance, Epidemiology, and End Results (SEER) 18 registry (2000 to 2011), SEER-Medicare linked files, and the National Health Interview Survey (NHIS) (1986 to 2009). Statistical methods developed to calculate natural life expectancy in the absence of the cancer, cancer-specific survival, and other-cause survival were applied to oral cancer data and internally validated with 10-fold cross-validation. Eligible participants were aged between 20 and 94 years with oral squamous cell carcinoma.

Exposures

Histologically confirmed oral cancer, general health status, smoking, and selected serious comorbid conditions.

Main Outcomes and Measures

Probabilities of surviving or dying from the cancer or from other causes, and life expectancy in the absence of the cancer.

Results

A total of 22 392 patients with oral squamous cell carcinoma (13 544 male [60.5%]; 1476 Asian and Pacific Islander [6.7%]; 1792 Black [8.0%], 1589 Hispanic [7.2%], 17 300 White [78.1%]) and 402 626 NHIS interviewees were included in this calculator designed for public use for patients ages 20 to 86 years with newly diagnosed oral cancer to obtain estimates of health status-adjusted age, life expectancy in the absence of the cancer, and the probability of surviving, dying from the cancer, or dying from other causes within 1 to 10 years after diagnosis. The models in the calculator estimated that patients with oral cancer have a higher risk of death from other causes than their matched US population, and that this risk increases by stage.

Conclusions and relevance

The models developed for the calculator demonstrate that survival estimates that exclude the effects of coexisting conditions can lead to underestimates or overestimates of survival. This new calculator approach will be broadly applicable for developing future prognostic models of cancer and noncancer aspects of a person's health in other cancers; as registries develop more linkages, available covariates will become broader, strengthening future tools.

myBeST-A Web-Based Survival Prognostic Tool for Women with Breast Cancer in Malaysia: Development Process and Preliminary Validation Study

Women with breast cancer are keen to know their predicted survival. We developed a new prognostic model for women with breast cancer in Malaysia. Using the model, this study aimed to design the user interface and develop the contents of a web-based prognostic tool for the care provider to convey survival estimates. We employed an iterative website development process which includes: (1) an initial development stage informed by reviewing existing tools and deliberation among breast surgeons and epidemiologists, (2) content validation and feedback by medical specialists, and (3) face validation and end-user feedback among medical officers. Several iterative prototypes were produced and improved based on the feedback. The experts (n = 8) highly agreed on the website content and predictors for survival with content validity indices ≥ 0.88. Users (n = 20) scored face validity indices of more than 0.90. They expressed favourable responses. The tool, named Malaysian Breast cancer Survival prognostic Tool (myBeST), is accessible online. The tool estimates an individualised five-year survival prediction probability. Accompanying contents were included to explain the tool's aim, target user, and development process. The tool could act as an additional tool to provide evidence-based and personalised breast cancer outcomes.

Predictive risk-score model to select patients with intrahepatic cholangiocarcinoma for adjuvant chemotherapy

BACKGROUND

The aim of this study was to develop a predictive model to identify individuals most likely to derive overall survival (OS) benefit from adjuvant chemotherapy (AC) after hepatic resection of intrahepatic cholangiocarcinoma (ICC).

METHODS

Patients who underwent hepatic resection of ICC between 1990 and 2020 were identified from a multi-institutional database. Factors associated with worse OS were identified and incorporated into an online predictive model to identify patients most likely to benefit from AC.

RESULTS

Among 726 patients, 189 (26.0%) individuals received AC. Factors associated with OS on multivariable analysis included CA19-9 (Hazard Ratio [HR]1.17, 95%CI 1.04-1.31), tumor burden score (HR1.09, 95%CI 1.04-1.15), T-category (T2/3/4, HR1.73, 95%CI 1.73-2.64), nodal disease (N1, HR3.80, 95%CI 2.02-7.15), tumor grade (HR1.88, 95%CI 1.00-3.55), and morphological subtype (HR2.19, 95%CI 1.08-4.46). A weighted predictive score was devised and made available online (https://yutaka-endo.shinyapps.io/ICCrisk_model_for_AC/). Receipt of AC was associated with a survival benefit among patients at high/medium-risk (high: no AC, 0% vs. AC, 20.6%; medium: no AC, 36.4% vs. 40.8%; both p < 0.05) but not low-risk (low: no AC, 65.1% vs. AC, 65.1%; p = 0.73) tumors.

CONCLUSION

An online predictive model based on tumor characteristics may help identify which patients may benefit the most from AC following resection of ICC.

A Novel Online Calculator to Predict Risk of Microvascular Invasion in the Preoperative Setting for Hepatocellular Carcinoma Patients Undergoing Curative-Intent Surgery

BACKGROUND

The presence of microvascular invasion (MVI) has been highlighted as an important determinant of hepatocellular carcinoma (HCC) prognosis. We sought to build and validate a novel model to predict MVI in the preoperative setting.

METHODS

Patients who underwent curative-intent surgery for HCC between 2000 and 2020 were identified using a multi-institutional database. Preoperative predictive models for MVI were built, validated, and used to develop a web-based calculator.

RESULTS

Among 689 patients, MVI was observed in 323 patients (46.9%). On multivariate analysis in the test cohort, preoperative parameters associated with MVI included α-fetoprotein (AFP; odds ratio [OR] 1.50, 95% confidence interval [CI] 1.23-1.83), imaging tumor burden score (TBS; hazard ratio [HR] 1.11, 95% CI 1.04-1.18), and neutrophil-to-lymphocyte ratio (NLR; OR 1.18, 95% CI 1.03-1.35). An online calculator to predict MVI was developed based on the weighted β-coefficients of these three variables ( https://yutaka-endo.shinyapps.io/MVIrisk/ ). The c-index of the test and validation cohorts was 0.71 and 0.72, respectively. Patients with a high risk of MVI had worse disease-free survival (DFS) and overall survival (OS) compared with low-risk MVI patients (3-year DFS: 33.0% vs. 51.9%, p < 0.001; 5-year OS: 44.2% vs. 64.8%, p < 0.001). DFS was worse among patients who underwent an R1 versus R0 resection among those patients at high risk of MVI (R0 vs. R1 resection: 3-year DFS, 36.3% vs. 16.1%, p = 0.002). In contrast, DFS was comparable among patients at low risk of MVI regardless of margin status (R0 vs. R1 resection: 3-year DFS, 52.9% vs. 47.3%, p = 0.16).

CONCLUSION

Preoperative assessment of MVI using the online tool demonstrated very good accuracy to predict MVI.

Integrated analysis of the clinical consequence and associated gene expression of ALK in ALK-positive human cancers

Anaplastic lymphoma kinase (ALK) is a tyrosine kinase receptor that is genetically altered in several cancers, including NSCLC, melanoma, lymphoma, and other tumors. Although ALK is associated with various cancers, the relationship between ALK expression and patient prognosis in different cancers is poorly understood. Here, using multidimensional approaches, we revealed the correlation between ALK expression and the clinical outcomes of patients with LUAD, melanoma, OV, DLBC, AML, and BC. We analyzed ALK transcriptional expression, patient survival rate, genetic alteration, protein network, and gene and microRNA (miRNA) co-expression. Compared to that in normal tissues, higher ALK expression was found in LUAD, melanoma, and OV, which are associated with poor patient survival rates. In contrast, lower transcriptional expression was found to decrease the survival rate of patients with DLBC, AML, and BC. A total of 202 missense mutations, 17 truncating mutations, 7 fusions, and 3 in-frame mutations were identified. Further, 17 genes and 19 miRNAs were found to be exclusively co-expressed and echinoderm microtubule-associated protein-like 4 (EML4) was identified as the most positively correlated gene (log odds ratio >3). The gene ontology and signaling pathways of the genes co-expressed with ALK in these six cancers were also identified. Our findings offer a basis for ALK as a prognostic biomarker and therapeutic target in cancers, which will potentially contribute to precision oncology and assist clinicians in identifying suitable treatment options.

Patient-centered outcomes for gastrointestinal cancer care: a scoping review protocol

INTRODUCTION

Following a cancer diagnosis, patients and their caregivers face crucial decisions regarding goals of care and treatment, which have consequences that can persist throughout their cancer journey. To foster informed and value-driven treatment choices, evidence-based information on outcomes relevant to patients is needed. Traditionally, clinical studies have largely focused on a few concrete and easily measurable outcomes such as survival, disease progression and immediate treatment toxicities. These outcomes do not capture other important factors that patients consider when making treatment decisions. Patient-centred outcomes (PCOs) reflect the patients' individual values, preferences, needs and circumstances that are essential to directing meaningful and informed healthcare discussions. Often, however, these outcomes are not included in research protocols in a standardised and practical fashion. This scoping review will summarise the existing literature on PCOs in gastrointestinal (GI) cancer care as well as the tools used to assess these outcomes. A comprehensive list of these PCOs will be generated for future efforts to develop a core outcome set.

METHODS AND ANALYSIS

This scoping review will follow Arksey and O'Malley's expanded framework for scoping reviews. We will systematically search Medline, Embase, CINAHL, Cochrane Library and APA PsycINFO databases for studies examining PCOs in the context of GI cancer. We will include studies published in or after the year 2000 up to the date of the final searches, with no language restrictions. Studies involving adult patients with GI cancers and discussion of any PCOs will be included. Opinion pieces, protocols, case reports and abstracts will be excluded. Two authors will independently perform two rounds of screening to select studies for inclusion. The data from full texts will be extracted, charted and summarised both quantitatively and qualitatively.

ETHICS AND DISSEMINATION

No ethics approval is required for this scoping review. Results will be disseminated through scientific publication and presentation at relevant conferences.

Prognostic disclosure in oncology - current communication models: a scoping review

BACKGROUND

Prognostic disclosure is essential to informed decision making in oncology, yet many oncologists are unsure how to successfully facilitate this discussion. This scoping review determines what prognostic communication models exist, compares and contrasts these models, and explores the supporting evidence.

METHOD

A protocol was created for this study using the Preferred Reporting Items for Systematic Reviews and Meta-analyses Protocols extension for Scoping Reviews. Comprehensive literature searches of electronic databases MEDLINE, EMBASE, PsycINFO and Cochrane CENTRAL were executed to identify relevant publications between 1971 and 2020.

RESULTS

In total, 1532 articles were identified, of which 78 met inclusion criteria and contained 5 communication models. Three of these have been validated in randomised controlled trials (the Serious Illness Conversation Guide, the Four Habits Model and the ADAPT acronym) and have demonstrated improved objective communication measures and patient reported outcomes. All three models emphasise the importance of exploring patients' illness understanding and treatment preferences, communicating prognosis and responding to emotion.

CONCLUSION

Communicating prognostic estimates is a core competency skill in advanced cancer care. This scoping review highlights available communication models and identifies areas in need of further assessment. Such areas include how to maintain learnt communication skills for lifelong practice, how to assess patient and caregiver understanding during and after these conversations, and how to best scale these protocols at the institutional and national levels.

A scoping review of interactive and personalized web-based clinical tools to support treatment decision making in breast cancer

The increasing attention on personalized breast cancer care has resulted in an explosion of new interactive, tailored, web-based clinical decision tools for guiding treatment decisions in clinical practice. The goal of this study was to review, compare, and discuss the clinical implications of current tools, and highlight future directions for tools aiming to improve personalized breast cancer care. We searched PubMed, Embase, PsychInfo, Cochrane Database of Systematic Reviews, Web of Science, and Scopus to identify web-based decision tools addressing breast cancer treatment decisions. There was a total of 17 articles associated with 21 unique tools supporting decisions related to surgery, radiation therapy, hormonal therapy, bisphosphonates, HER2-targeted therapy, and chemotherapy. The quality of the tools was assessed using the International Patient Decision Aid Standard instrument. Overall, the tools considered clinical (e.g., age) and tumor characteristics (e.g., grade) to provide personalized outcomes (e.g., survival) associated with various treatment options. Fewer tools provided the adverse effects of the selected treatment. Only one tool was field-tested with patients, and none were tested with healthcare providers. Future studies need to assess the feasibility, usability, acceptability, as well as the effects of personalized web-based decision tools on communication and decision making from the patient and clinician perspectives.

Exploring Cancer Survivor Needs and Preferences for Communicating Personalized Cancer Statistics From Registry Data: Qualitative Multimethod Study

Background

Disclosure of cancer statistics (eg, survival or incidence rates) based on a representative group of patients can help increase cancer survivors’ understanding of their own diagnostic and prognostic situation, and care planning. More recently, there has been an increasing interest in the use of cancer registry data for disclosing and communicating personalized cancer statistics (tailored toward personal and clinical characteristics) to cancer survivors and relatives.

Objective

The aim of this study was to explore breast cancer (BCa) and prostate cancer (PCa) survivor needs and preferences for disclosing (what) and presenting (how) personalized statistics from a large Dutch population-based data set, the Netherlands Cancer Registry (NCR).

Methods

To elicit survivor needs and preferences for communicating personalized NCR statistics, we created different (non)interactive tools visualizing hypothetical scenarios and adopted a qualitative multimethod study design. We first conducted 2 focus groups (study 1; n=13) for collecting group data on BCa and PCa survivor needs and preferences, using noninteractive sketches of what a tool for communicating personalized statistics might look like. Based on these insights, we designed a revised interactive tool, which was used to further explore the needs and preferences of another group of cancer survivors during individual think-aloud observations and semistructured interviews (study 2; n=11). All sessions were audio-recorded, transcribed verbatim, analyzed using thematic (focus groups) and content analysis (think-aloud observations), and reported in compliance with qualitative research reporting criteria.

Results

In both studies, cancer survivors expressed the need to receive personalized statistics from a representative source, with especially a need for survival and conditional survival rates (ie, survival rate for those who have already survived for a certain period). Personalized statistics adjusted toward personal and clinical factors were deemed more relevant and useful to know than generic or average-based statistics. Participants also needed support for correctly interpreting the personalized statistics and putting them into perspective, for instance by adding contextual or comparative information. Furthermore, while thinking aloud, participants experienced a mix of positive (sense of hope) and negative emotions (feelings of distress) while viewing the personalized survival data. Overall, participants preferred simplicity and conciseness, and the ability to tailor the type of visualization and amount of (detailed) statistical information.

Conclusions

The majority of our sample of cancer survivors wanted to receive personalized statistics from the NCR. Given the variation in patient needs and preferences for presenting personalized statistics, designers of similar information tools may consider potential tailoring strategies on multiple levels, as well as effective ways for providing supporting information to make sure that the personalized statistics are properly understood. This is encouraging for cancer registries to address this unmet need, but also for those who are developing or implementing personalized data-driven information tools for patients and relatives.

Development and validation of a prediction model of poor performance status and severe symptoms over time in cancer patients (PROVIEW+)

BACKGROUND

Predictive cancer tools focus on survival; none predict severe symptoms.

AIM

To develop and validate a model that predicts the risk for having low performance status and severe symptoms in cancer patients.

DESIGN

Retrospective, population-based, predictive study.

SETTING/PARTICIPANTS

We linked administrative data from cancer patients from 2008 to 2015 in Ontario, Canada. Patients were randomly selected for model derivation (60%) and validation (40%). Using the derivation cohort, we developed a multivariable logistic regression model to predict the risk of an outcome at 6 months following diagnosis and recalculated after each of four annual survivor marks. Model performance was assessed using discrimination and calibration plots. Outcomes included low performance status (i.e. 10-30 on Palliative Performance Scale), severe pain, dyspnea, well-being, and depression (i.e. 7-10 on Edmonton Symptom Assessment System).

RESULTS

We identified 255,494 cancer patients (57% female; median age of 64; common cancers were breast (24%); and lung (13%)). At diagnosis, the predicted risk of having low performance status, severe pain, well-being, dyspnea, and depression in 6-months is 1%, 3%, 6%, 13%, and 4%, respectively for the reference case (i.e. male, lung cancer, stage I, no symptoms); the corresponding discrimination for each outcome model had high AUCs of 0.807, 0.713, 0.709, 0.790, and 0.723, respectively. Generally these covariates increased the outcome risk by >10% across all models: lung disease, dementia, diabetes; radiation treatment; hospital admission; pain; depression; transitional performance status; issues with appetite; or homecare.

CONCLUSIONS

The model accurately predicted changing cancer risk for low performance status and severe symptoms over time.

Interactive Data Visualization Tool for Patient-Centered Decision Making in Kidney Cancer

PURPOSE

Patients and providers often lack clinical decision tools to enable effective shared decision making. This is especially true in the rapidly changing therapeutic landscape of metastatic kidney cancer. Using the International Metastatic Renal Cell Carcinoma Database Consortium (IMDC) criteria, a validated risk prediction tool for patients with metastatic renal cell carcinoma, we created and user-tested a novel interactive visualization for clinical use.

METHODS

An interactive visualization depicting IMDC criteria was created, with the final version including data for more than 4,500 patients. Usability testing was performed with nonmedical lay-users and medical oncology fellow physicians. Subjects used the tool to calculate median survival times based on IMDC criteria. User confidence was surveyed. An iterative user feedback implementation cycle was completed and informed revision of the tool.

RESULTS

The tool is available at CloViz-IMDC. Initially, 400 lay-users and 15 physicians completed clinical scenarios and surveys. Cumulative accuracy across scenarios was higher for physicians than lay-users (84% v 74%; P = .03). Eighty-three percent of lay-users and 87% of physicians thought the tool became intuitive with use. Sixty-eight percent of lay-users wanted to use the tool clinically compared with 87% of physicians. After revisions, the updated tool was user-tested with 100 lay-users and 15 physicians. Physicians, but not lay-users, showed significant improvement in accuracy in the updated version of the tool (90% v 67%; P = .008). Seventy-two percent of lay-users and 93% of physicians wanted to use the updated tool in a clinical setting.

CONCLUSION

A graphical method of interacting with a validated nomogram provides prognosis results that can be used by nonmedical lay-users and physicians, and has the potential for expanded use across many clinical conditions.

Development and Assessment of a Model for Predicting Individualized Outcomes in Patients With Oropharyngeal Cancer

IMPORTANCE

Recent insights into the biologic characteristics and treatment of oropharyngeal cancer may help inform improvements in prognostic modeling. A bayesian multistate model incorporates sophisticated statistical techniques to provide individualized predictions of survival and recurrence outcomes for patients with newly diagnosed oropharyngeal cancer.

OBJECTIVE

To develop a model for individualized survival, locoregional recurrence, and distant metastasis prognostication for patients with newly diagnosed oropharyngeal cancer, incorporating clinical, oncologic, and imaging data.

DESIGN, SETTING, AND PARTICIPANTS

In this prognostic study, a data set was used comprising 840 patients with newly diagnosed oropharyngeal cancer treated at a National Cancer Institute-designated center between January 2003 and August 2016; analysis was performed between January 2019 and June 2020. Using these data, a bayesian multistate model was developed that can be used to obtain individualized predictions. The prognostic performance of the model was validated using data from 447 patients treated for oropharyngeal cancer at Erasmus Medical Center in the Netherlands.

EXPOSURES

Clinical/oncologic factors and imaging biomarkers collected at or before initiation of first-line therapy.

MAIN OUTCOMES AND MEASURES

Overall survival, locoregional recurrence, and distant metastasis after first-line cancer treatment.

RESULTS

Of the 840 patients included in the National Cancer Institute-designated center, 715 (85.1%) were men and 268 (31.9%) were current smokers. The Erasmus Medical Center cohort comprised 300 (67.1%) men, with 350 (78.3%) current smokers. Model predictions for 5-year overall survival demonstrated good discrimination, with area under the curve values of 0.81 for the model with and 0.78 for the model without imaging variables. Application of the model without imaging data in the independent Dutch validation cohort resulted in an area under the curve of 0.75. This model possesses good calibration and stratifies patients well in terms of likely outcomes among many competing events.

CONCLUSIONS AND RELEVANCE

In this prognostic study, a multistate model of oropharyngeal cancer incorporating imaging biomarkers appeared to estimate and discriminate locoregional recurrence from distant metastases. Providing personalized predictions of multiple outcomes increases the information available for patients and clinicians. The web-based application designed in this study may serve as a useful tool for generating predictions and visualizing likely outcomes for a specific patient.

Cancer survival statistics for patients and healthcare professionals - a tutorial of real-world data analysis

Monitoring survival of cancer patients using data collected by population-based cancer registries is an important component of cancer control. In this setting, patient survival is often summarized using net survival, that is survival from cancer if there were no other possible causes of death. Although net survival is the gold standard for comparing survival between groups or over time, it is less relevant for understanding the anticipated real-world prognosis of patients. In this review, we explain statistical concepts targeted towards patients, clinicians and healthcare professionals that summarize cancer patient survival under the assumption that other causes of death exist. Specifically, we explain the appropriate use, interpretation and assumptions behind statistical methods for competing risks, loss in life expectancy due to cancer and conditional survival. These concepts are relevant when producing statistics for risk communication between physicians and patients, planning for use of healthcare resources, or other applications when consideration of both cancer outcomes and the competing risks of death is required. To reinforce the concepts, we use Swedish population-based data of patients diagnosed with cancer of the breast, prostate, colon and chronic myeloid leukaemia. We conclude that when choosing between summary measures of survival it is critical to characterize the purpose of the study and to determine the nature of the hypothesis under investigation. The choice of terminology and style of reporting should be carefully adapted to the target audience and may range from summaries for specialist readers of scientific publications to interactive online tools aimed towards lay persons.

Stage IIIa Melanoma and Impact of Multiple Positive Lymph Nodes on Survival

BACKGROUND

For patients with cutaneous melanoma, having >1 positive lymph node (LN) is associated with worse survival. We hypothesized that for stage IIIA patients, N2a disease (2 to 3 positive LN) would be associated with a worse prognosis compared to those with N1a disease (1 positive LN).

STUDY DESIGN

Stage IIIA melanoma patients in the NCDB Participant User File from 2010 to 2016 were analyzed. Overall survival (OS) between N1a and N2a patients was compared. Subgroup analyses were made between patients undergoing sentinel lymph node (SLN) biopsy alone and those undergoing subsequent completion lymph node dissection (CLND). A separate post hoc analysis of T2a patients undergoing SLN biopsy and CLND from a prospective multicenter randomized clinical trial was performed to validate the findings.

RESULTS

Records of 2,305 IIIA patients were evaluated. In an adjusted survival model, N2a disease was an independent risk factor for worse OS (hazard ratio [HR] 1.56, p = 0.0052). In the subgroup analysis, there was no difference in OS between N1a and N2a disease for patients who underwent SLN biopsy without CLND (p = 0.59), but there was a significant difference in OS for patients who underwent SLN biopsy plus CLND (p = 0.0009). The separate clinical trial database confirmed that for patients with SLN-only disease, there was no difference in OS between N1a and N2a disease.

CONCLUSIONS

For stage IIIA melanoma patients, the distribution of micrometastatic lymph node disease (SLN or non-SLN), rather than the absolute number of SLNs, should be considered when individualizing adjuvant therapy recommendations.

Logistic regression has similar performance to optimised machine learning algorithms in a clinical setting: application to the discrimination between type 1 and type 2 diabetes in young adults

BACKGROUND

There is much interest in the use of prognostic and diagnostic prediction models in all areas of clinical medicine. The use of machine learning to improve prognostic and diagnostic accuracy in this area has been increasing at the expense of classic statistical models. Previous studies have compared performance between these two approaches but their findings are inconsistent and many have limitations. We aimed to compare the discrimination and calibration of seven models built using logistic regression and optimised machine learning algorithms in a clinical setting, where the number of potential predictors is often limited, and externally validate the models.

METHODS

We trained models using logistic regression and six commonly used machine learning algorithms to predict if a patient diagnosed with diabetes has type 1 diabetes (versus type 2 diabetes). We used seven predictor variables (age, BMI, GADA islet-autoantibodies, sex, total cholesterol, HDL cholesterol and triglyceride) using a UK cohort of adult participants (aged 18-50 years) with clinically diagnosed diabetes recruited from primary and secondary care (n = 960, 14% with type 1 diabetes). Discrimination performance (ROC AUC), calibration and decision curve analysis of each approach was compared in a separate external validation dataset (n = 504, 21% with type 1 diabetes).

RESULTS

Average performance obtained in internal validation was similar in all models (ROC AUC ≥ 0.94). In external validation, there were very modest reductions in discrimination with AUC ROC remaining ≥ 0.93 for all methods. Logistic regression had the numerically highest value in external validation (ROC AUC 0.95). Logistic regression had good performance in terms of calibration and decision curve analysis. Neural network and gradient boosting machine had the best calibration performance. Both logistic regression and support vector machine had good decision curve analysis for clinical useful threshold probabilities.

CONCLUSION

Logistic regression performed as well as optimised machine algorithms to classify patients with type 1 and type 2 diabetes. This study highlights the utility of comparing traditional regression modelling to machine learning, particularly when using a small number of well understood, strong predictor variables.

Development and Validation of a Prognostic Survival Model With Patient-Reported Outcomes for Patients With Cancer

IMPORTANCE

Existing prognostic cancer tools include biological and laboratory variables. However, patients often do not know this information, preventing them from using the tools and understanding their prognosis.

OBJECTIVE

To develop and validate a prognostic survival model for all cancer types that incorporates information on symptoms and performance status over time.

DESIGN, SETTING, AND PARTICIPANTS

This is a retrospective, population-based, prognostic study of data from patients diagnosed with cancer from January 1, 2008, to December 31, 2015, in Ontario, Canada. Patients were randomly selected for model derivation (60%) and validation (40%). The derivation cohort was used to develop a multivariable Cox proportional hazards regression model with baseline characteristics under a backward stepwise variable selection process to predict the risk of mortality as a function of time. Covariates included demographic characteristics, clinical information, symptoms and performance status, and health care use. Model performance was assessed on the validation cohort by C statistics and calibration plots. Data analysis was performed from February 6, 2018, to November 6, 2019.

MAIN OUTCOMES AND MEASURES

Time to death from diagnosis (year 0) recalculated at each of 4 annual survivor marks after diagnosis (up to year 4).

RESULTS

A total of 255 494 patients diagnosed with cancer were identified (135 699 [53.1%] female; median age, 65 years [interquartile range, 55-73 years]). The cohort decreased to 217 055, 184 822, 143 649, and 109 569 patients for each of the 4 years after diagnosis. In the derivation cohort year 0, and the most common cancers were breast (30 855 [20.1%]), lung (19 111 [12.5%]), and prostate (18 404 [12.0%]). A total of 47 614 (31.1%) had stage III or IV disease. The mean (SD) time to death in year 0 was 567 (715) days. After backward stepwise selection in year 0, the following factors were associated with increased risk of death by more than 10%: being hospitalized; having congestive heart failure, chronic obstructive pulmonary disease, or dementia; having moderate to high pain; having worse well-being; having functional status in the transitional or end-of-life phase; having any problems with appetite; receiving end-of-life home care; and living in a nursing home. Model discrimination was high for all models (C statistic: 0.902 [year 0], 0.912 [year 1], 0.912 [year 2], 0.909 [year 3], and 0.908 [year 4]).

CONCLUSIONS AND RELEVANCE

The model accurately predicted changing cancer survival risk over time using clinical, symptom, and performance status data and appears to have the potential to be a useful prognostic tool that can be completed by patients. This knowledge may support earlier integration of palliative care.

Using the Technology Acceptance Model to Develop StartSmart: mHealth for Screening, Brief Intervention, and Referral for Risk and Protective Factors in Pregnancy

INTRODUCTION

Technology decision support with tailored patient education has the potential to improve maternal and child health outcomes. The purpose of this study was to develop StartSmart, a mobile health (mHealth) intervention to support evidence-based prenatal screening, brief intervention, and referral to treatment for risk and protective factors in pregnancy.

METHODS

StartSmart was developed using Davis' Technology Acceptance Model with end users engaged in the technology development from initial concept to clinical testing. The prototype was developed based upon the current guidelines, focus group findings, and consultation with patient and provider experts. The prototype was then alpha tested by clinicians and patients. Clinicians were asked to give feedback on the screening questions, treatment, brief motivational interviewing, referral algorithms, and the individualized education materials. Clinicians were asked about the feasibility of using the materials to provide brief intervention or referral to treatment. Patients were interviewed using the think aloud technique, a cognitive engineering method used to inform the design of mHealth interventions. Interview questions were guided by the Screening, Brief Intervention, Referral to Treatment theory and attention to usefulness and usability.

RESULTS

Expert clinicians provided guidance on the screening instruments, resources, and practice guidelines. Clinicians suggested identifying specific prenatal visits for the screening (first prenatal visit, 28-week visit, and 36-week visit). Patients reported that the tablet-based screening was useful to promote adherence to guidelines and provided suggestions for improvement including more information on the diabetic diet and more resources for diabetes. During alpha testing, participants commented on navigability and usability. Patients reported favorable responses about question wording and ease of use.

DISCUSSION

Clinicians reported the use of mHealth to screen and counsel pregnant patients on risk and protective factors facilitated their ability to provide comprehensive care.

On the validity of time-dependent AUC estimation in the presence of cure fraction

During the last decades, several approaches have been proposed to estimate the time-dependent area under the receiver operating characteristic curve (AUC) of risk tools derived from survival data. The validity of these estimators relies on some regularity assumptions among which a survival function being proper. In practice, this assumption is not always satisfied because a fraction of the population may not be susceptible to experience the event of interest even for long follow-up. Studying the sensitivity of the proposed estimators to the violation of this assumption is of substantial interest. In this paper, we investigate the performance of a nonparametric simple estimator, developed for classical survival data, in the case when the population exhibits a cure fraction. Motivated from the current practice of deriving risk tools in oncology and cardiovascular disease prevention, we also assess the loss, in terms of predictive performance, when deriving risk tools from survival models that do not acknowledge the presence of cure. The simulation results show that the investigated method is valid even under the presence of cure. They also show that risk tools derived from survival models that ignore the presence of cure have smaller AUC compared to those derived from survival models that acknowledge the presence of cure. This was also attested with a real data analysis from a breast cancer study.

Opportunities for personalised follow‐up care among patients with breast cancer: A scoping review to identify preference‐sensitive decisions

Introduction

Current follow‐up arrangements for breast cancer do not optimally meet the needs of individual patients. We therefore reviewed the evidence on preferences and patient involvement in decisions about breast cancer follow‐up to explore the potential for personalised care.

Methods

Studies published between 2008 and 2017 were extracted from MEDLINE, PsycINFO and EMBASE. We then identified decision categories related to content and form of follow‐up. Criteria for preference sensitiveness and patient involvement were compiled and applied to determine the extent to which decisions were sensitive to patient preferences and patients were involved.

Results

Forty‐one studies were included in the full‐text analysis. Four decision categories were identified: “surveillance for recurrent/secondary breast cancer; consultations for physical and psychosocial effects; recurrence‐risk reduction by anti‐hormonal treatment; and improving quality of life after breast cancer.” There was little evidence that physicians treated decisions about anti‐hormonal treatment, menopausal symptoms, and follow‐up consultations as sensitive to patient preferences. Decisions about breast reconstruction were considered as very sensitive to patient preferences, and patients were usually involved.

Conclusion

Patients are currently not involved in all decisions that affect them during follow‐up, indicating a need for improvements. Personalised follow‐up care could improve resource allocation and the value of care for patients.

Communicating uncertainty in cancer prognosis: A review of web-based prognostic tools

Objective To review how web-based prognosis tools for cancer patients and clinicians describe aleatory (risk estimates) and epistemic (imprecision in risk estimates) uncertainties. Methods We reviewed prognostic tools available online and extracted all uncertainty descriptions. We adapted an existing classification and classified each extracted statement by presentation of uncertainty. Results We reviewed 222 different prognostic risk tools, which produced 772 individual estimates. When describing aleatory uncertainty, almost all (90%) prognostic tools included a quantitative description, such as "chances of survival after surgery are 10%", though there was heterogeneity in the use of percentages, natural frequencies, and use of graphics. Only 14% of tools described epistemic uncertainty. Of those that did, most used a qualitative prefix such as "about" or "up to", while 22 tools described quantitative descriptions using confidence intervals or ranges. Conclusions Considerable heterogeneity exists in the way uncertainties are communicated in cancer prognostic tools. Few tools describe epistemic uncertainty. This variation is predominately explained by a lack of evidence and consensus in risk communication, particularly for epistemic uncertainty. Practice Implications As precision medicine seeks to improve prognostic estimates, the community may not be equipped with the tools to communicate the results accurately and effectively to clinicians and patients.

A novel online prognostic tool to predict long-term survival after liver resection for intrahepatic cholangiocarcinoma: The "metro-ticket" paradigm

BACKGROUND

The aim of the current study was to develop an online calculator to predict survival after liver resection for intrahepatic cholangiocarcinoma (ICC) based on the "metro-ticket" paradigm.

METHODS

Between 1990 and 2016, patients who underwent liver resection for ICC were identified in an international multi-institutional database. The final multivariable model of survival was used to develop an online prognostic calculator of survival.

RESULTS

Among 643 patients, actual 5-year overall survival (OS) after resection for ICC was 42.7%. On multivariable analysis, CA19-9 > 200 (hazard ratio (HR), 2.62; 95% CI, 2.01-3.42), sum of the number and largest tumor size >7 (HR, 1.88; 95% CI, 1.46-2.42), N1 disease (HR, 2.87; 95% CI, 1.98-4.16), R1 resection (HR, 1.72; 95% CI, 1.21-2.46), poor/undifferentiated tumor grade (HR, 1.74; 95% CI, 1.25-2.44), major vascular invasion (HR, 1.47; 95% CI, 1.03-2.10), and adjuvant chemotherapy (HR, 0.64; 95% CI, 0.45-0.89) were significantly associated with survival and were included in the online calculator. The predictive accuracy of the model was good to very good as the C-statistics to predict 5-year OS was 0.696 in the training dataset and 0.672 with bootstrapping resamples (n = 5000) in the test dataset.

CONCLUSION

A novel, online calculator was developed to estimate the 5-year survival probability for patients undergoing resection for ICC. This tool could help provide useful information to guide treatment decision-making and inform conversations about prognosis.

Communicating prognosis to women with early breast cancer - overview of prediction tools and the development and pilot testing of a decision aid

Background

Shared decision-making in oncology requires information on individual prognosis. This comprises cancer prognosis as well as competing risks of dying due to age and comorbidities. Decision aids usually do not provide such information on competing risks. We conducted an overview on clinical prediction tools for early breast cancer and developed and pilot-tested a decision aid (DA) addressing individual prognosis using additional chemotherapy in early, hormone receptor-positive breast cancer as an example.

Methods

Systematic literature search on clinical prediction tools for the effects of drug treatment on survival of breast cancer. The DA was developed following criteria for evidence-based patient information and International Patient Decision Aids Standards. We included data on the influence of age and comorbidities on overall prognosis. The DA was pilot-tested in focus groups. Comprehension was additionally evaluated through an online survey with women in breast cancer self-help groups.

Results

We identified three prediction tools: Adjuvant!Online, PREDICT and CancerMath. All tools consider age and tumor characteristics. Adjuvant!Online considers comorbidities, CancerMath displays age-dependent non-cancer mortality. Harm due to therapy is not reported.

Twenty women participated in focus groups piloting the DA until data saturation was achieved. A total of 102 women consented to participate in the online survey, of which 86 completed the survey. The rate of correct responses was 90.5% and ranged between 84 and 95% for individual questions.

Conclusions

None of the clinical prediction tools fulfilled the requirements to provide women with all the necessary information for informed decision-making. Information on individual prognosis was well understood and can be included in patient decision aids.

Electronic supplementary material

The online version of this article (10.1186/s12913-019-3988-2) contains supplementary material, which is available to authorized users.

Individualized survival prediction for patients with oropharyngeal cancer in the human papillomavirus era

BACKGROUND

Accurate, individualized prognostication in patients with oropharyngeal squamous cell carcinoma (OPSCC) is vital for patient counseling and treatment decision making. With the emergence of human papillomavirus (HPV) as an important biomarker in OPSCC, calculators incorporating this variable have been developed. However, it is critical to characterize their accuracy prior to implementation.

METHODS

Four OPSCC calculators were identified that integrate HPV into their estimation of 5-year overall survival. Treatment outcomes for 856 patients with OPSCC who were evaluated at a single institution from 2003 through 2016 were analyzed. Predicted survival probabilities were generated for each patient using each calculator. Calculator performance was assessed and compared using Kaplan-Meier plots, receiver operating characteristic curves, concordance statistics, and calibration plots.

RESULTS

Correlation between pairs of calculators varied, with coefficients ranging from 0.63 to 0.90. Only 3 of 6 pairs of calculators yielded predictions within 10% of each other for at least 50% of patients. Kaplan-Meier curves of calculator-defined risk groups demonstrated reasonable stratification. Areas under the receiver operating characteristic curve ranged from 0.74 to 0.80, and concordance statistics ranged from 0.71 to 0.78. Each calculator demonstrated superior discriminatory ability compared with clinical staging according to the seventh and eighth editions of the American Joint Committee on Cancer staging manual. Among models, the Denmark calculator was found to be best calibrated to observed outcomes.

CONCLUSIONS

Existing calculators exhibited reasonable estimation of survival in patients with OPSCC, but there was considerable variability in predictions for individual patients, which limits the clinical usefulness of these calculators. Given the increasing role of personalized treatment in patients with OPSCC, further work is needed to improve accuracy and precision, possibly through the identification and incorporation of additional biomarkers.

InterPreT cancer survival: A dynamic web interactive prediction cancer survival tool for health-care professionals and cancer epidemiologists

BACKGROUND

There are a variety of ways for quantifying cancer survival with each measure having advantages and disadvantages. Distinguishing these measures and how they should be interpreted has led to confusion among scientists, the media, health care professionals and patients. This motivates the development of tools to facilitate communication and interpretation of these statistics.

METHODS

"InterPreT Cancer Survival" is a newly developed, publicly available, online interactive cancer survival tool targeted towards health-care professionals and epidemiologists (http://interpret.le.ac.uk). It focuses on the correct interpretation of commonly reported cancer survival measures facilitated through the use of dynamic interactive graphics. Statistics presented are based on parameter estimates obtained from flexible parametric relative survival models using large population-based English registry data containing information on survival across 6 cancer sites; Breast, Colon, Rectum, Stomach, Melanoma and Lung.

RESULTS

Through interactivity, the tool improves understanding of various measures and how survival or mortality may vary by age and sex. Routine measures of cancer survival are reported, however, individualised estimates using crude probabilities are advocated, which is more appropriate for patients or health care professionals. The results are presented in various interactive formats facilitating understanding of individual risk and differences between various measures.

CONCLUSIONS

"InterPreT Cancer Survival" is presented as an educational tool which engages the user through interactive features to improve the understanding of commonly reported cancer survival statistics. The tool has received positive feedback from a Cancer Research UK patient sounding board and there are further plans to incorporate more disease characteristics, e.g. stage.

Variability in Predictions from Online Tools: A Demonstration Using Internet-Based Melanoma Predictors

BACKGROUND

Prognostic models are increasingly being made available online, where they can be publicly accessed by both patients and clinicians. These online tools are an important resource for patients to better understand their prognosis and for clinicians to make informed decisions about treatment and follow-up. The goal of this analysis was to highlight the possible variability in multiple online prognostic tools in a single disease.

METHODS

To demonstrate the variability in survival predictions across online prognostic tools, we applied a single validation dataset to three online melanoma prognostic tools. Data on melanoma patients treated at Memorial Sloan Kettering Cancer Center between 2000 and 2014 were retrospectively collected. Calibration was assessed using calibration plots and discrimination was assessed using the C-index.

RESULTS

In this demonstration project, we found important differences across the three models that led to variability in individual patients' predicted survival across the tools, especially in the lower range of predictions. In a validation test using a single-institution data set, calibration and discrimination varied across the three models.

CONCLUSIONS

This study underscores the potential variability both within and across online tools, and highlights the importance of using methodological rigor when developing a prognostic model that will be made publicly available online. The results also reinforce that careful development and thoughtful interpretation, including understanding a given tool's limitations, are required in order for online prognostic tools that provide survival predictions to be a useful resource for both patients and clinicians.

Predicting 6- and 12-Month Risk of Mortality in Patients With Platinum-Resistant Advanced-Stage Ovarian Cancer: Prognostic Model to Guide Palliative Care Referrals

OBJECTIVE

Predictive models are increasingly being used in clinical practice. The aim of the study was to develop a predictive model to identify patients with platinum-resistant ovarian cancer with a prognosis of less than 6 to 12 months who may benefit from immediate referral to hospice care.

METHODS

A retrospective chart review identified patients with platinum-resistant epithelial ovarian cancer who were treated at our institution between 2000 and 2011. A predictive model for survival was constructed based on the time from development of platinum resistance to death. Multivariate logistic regression modeling was used to identify significant survival predictors and to develop a predictive model. The following variables were included: time from diagnosis to platinum resistance, initial stage, debulking status, number of relapses, comorbidity score, albumin, hemoglobin, CA-125 levels, liver/lung metastasis, and the presence of a significant clinical event (SCE). An SCE was defined as a malignant bowel obstruction, pleural effusion, or ascites occurring on or before the diagnosis of platinum resistance.

RESULTS

One hundred sixty-four patients met inclusion criteria. In the regression analysis, only an SCE and the presence of liver or lung metastasis were associated with poorer short-term survival (P < 0.001). Nine percent of patients with an SCE or liver or lung metastasis survived 6 months or greater and 0% survived 12 months or greater, compared with 85% and 67% of patients without an SCE or liver or lung metastasis, respectively.

CONCLUSIONS

Patients with platinum-resistant ovarian cancer who have experienced an SCE or liver or lung metastasis have a high risk of death within 6 months and should be considered for immediate referral to hospice care.

Online tools for patient counseling in bladder and kidney cancer-ready for prime time?

Gauging prognosis is a key element when facing treatment decisions in cancer care. Several prognostic tools, such as risk tables and nomograms are at hand to aid this process. In the context of patient-centered care, prognostic tools are of great interest to caregivers and -providers alike, as they can convey sizeable amounts of information and provide tailored, accurate estimates of prognosis. Given the rising number of prognostic tools in cancer care over the last two decades, and similarly, ever increasing presence of the Internet, we aimed to assess how this would translate into the availability of online tools for patient counseling. We used a modified systematic review to evaluate the web-based availability, format, and content of prognostic tools for bladder and kidney cancer care. Our search identified a total of twenty-three tools, offered by eight providers, which assessed a total of six (bladder cancer) and five (kidney cancer) different outcomes. Despite the restricted availability of online tools, we observed that the majority showed limited user-friendliness (including, for example, a statement/explanation of intended use, visualization of data, availability as application software for handheld devices). Only one tool included modifiable risk factors such as smoking behavior and body weight. Lastly, none of the tools incorporated genomic or molecular markers or treatment associated quality of life. Taken together, online tools for patient counseling in bladder and kidney cancer care are only beginning to align with the growing need in clinical reality. Further and future avenues include incorporation of health-related quality of life as well as genomic and biomarkers into prediction tools.

Individualized outcome prognostication for patients with laryngeal cancer

BACKGROUND

Accurate prognostication is essential to the optimal management of laryngeal cancer. Predictive models have been developed to calculate the risk of oncologic outcomes, but extensive external validation of accuracy and reliability is necessary before implementing them into clinical practice.

METHOD

Four published prognostic calculators that predict 5-year overall survival for patients with laryngeal cancer were evaluated using patient information from a prospective epidemiology study cohort (n = 246; median follow-up, 60 months) with previously untreated, stage I through IVb laryngeal squamous cell carcinoma.

RESULTS

Different calculators yielded substantially different predictions for individual patients. The observed 5-year overall survival was significantly higher than the averaged predicted 5-year overall survival of the 4 calculators (71.9%; 95% confidence interval [CI], 65%-78%] vs 47.7%). Statistical analyses demonstrated the calculators' limited capacity to discriminate outcomes for risk-stratified patients. The area under the receiver operating characteristic curve ranged from 0.68 to 0.72. C-index values were similar for each of the 4 models (range, 0.66-0.68). There was a lower than expected hazard of death for patients who received induction (bioselective) chemotherapy (hazard ratio, 0.46; 95% CI, 0.24-0.88; P = .024) or primary surgical intervention (hazard ratio, 0.43; 95 % CI, 0.21-0.90; P = .024) compared with those who received concurrent chemoradiation.

CONCLUSIONS

Suboptimal reliability and accuracy limit the integration of existing individualized prediction tools into routine clinical decision making. The calculators predicted significantly worse than observed survival among patients who received induction chemotherapy and primary surgery, suggesting a need for updated consideration of modern treatment modalities. Further development of individualized prognostic calculators may improve risk prediction, treatment planning, and counseling for patients with laryngeal cancer. Cancer 2018;124:706-16. © 2017 American Cancer Society.

Development, Feasibility, and Small-Scale Implementation of a Web-Based Prognostic Tool-Surveillance, Epidemiology, and End Results Cancer Survival Calculator

BACKGROUND

Population datasets and the Internet are playing an ever-growing role in the way cancer information is made available to providers, patients, and their caregivers. The Surveillance, Epidemiology, and End Results Cancer Survival Calculator (SEER*CSC) is a Web-based cancer prognostic tool that uses SEER data, a large population dataset, to provide physicians with highly valid, evidence-based prognostic estimates for increasing shared decision-making and improving patient-provider communication of complex health information.

OBJECTIVE

The aim of this study was to develop, test, and implement SEER*CSC.

METHODS

An iterative approach was used to develop the SEER*CSC. Based on input from cancer patient advocacy groups and physicians, an initial version of the tool was developed. Next, providers from 4 health care delivery systems were recruited to do formal usability testing of SEER*CSC. A revised version of SEER*CSC was then implemented in two health care delivery sites using a real-world clinical implementation approach, and usage data were collected. Post-implementation follow-up interviews were conducted with site champions. Finally, patients from two cancer advocacy groups participated in usability testing.

RESULTS

Overall feedback of SEER*CSC from both providers and patients was positive, with providers noting that the tool was professional and reliable, and patients finding it to be informational and helpful to use when discussing their diagnosis with their provider. However, use during the small-scale implementation was low. Reasons for low usage included time to enter data, not having treatment options in the tool, and the tool not being incorporated into the electronic health record (EHR). Patients found the language in its current version to be too complex.

CONCLUSIONS

The implementation and usability results showed that participants were enthusiastic about the use and features of SEER*CSC, but sustained implementation in a real-world clinical setting faced significant challenges. As a result of these findings, SEER*CSC is being redesigned with more accessible language for a public facing release. Meta-tools, which put different tools in context of each other, are needed to assist in understanding the strengths and limitations of various tools and their place in the clinical decision-making pathway. The continued development and eventual release of prognostic tools should include feedback from multidisciplinary health care teams, various stakeholder groups, patients, and caregivers.

More Is Better: Recent Progress in Multi-Omics Data Integration Methods

Multi-omics data integration is one of the major challenges in the era of precision medicine. Considerable work has been done with the advent of high-throughput studies, which have enabled the data access for downstream analyses. To improve the clinical outcome prediction, a gamut of software tools has been developed. This review outlines the progress done in the field of multi-omics integration and comprehensive tools developed so far in this field. Further, we discuss the integration methods to predict patient survival at the end of the review.

PROACT: Iterative Design of a Patient-Centered Visualization for Effective Prostate Cancer Health Risk Communication

Prostate cancer is the most common cancer among men in the US, and yet most cases represent localized cancer for which the optimal treatment is unclear. Accumulating evidence suggests that the available treatment options, including surgery and conservative treatment, result in a similar prognosis for most men with localized prostate cancer. However, approximately 90% of patients choose surgery over conservative treatment, despite the risk of severe side effects like erectile dysfunction and incontinence. Recent medical research suggests that a key reason is the lack of patient-centered tools that can effectively communicate personalized risk information and enable them to make better health decisions. In this paper, we report the iterative design process and results of developing the PROgnosis Assessment for Conservative Treatment (PROACT) tool, a personalized health risk communication tool for localized prostate cancer patients. PROACT utilizes two published clinical prediction models to communicate the patients' personalized risk estimates and compare treatment options. In collaboration with the Maine Medical Center, we conducted two rounds of evaluations with prostate cancer survivors and urologists to identify the design elements and narrative structure that effectively facilitate patient comprehension under emotional distress. Our results indicate that visualization can be an effective means to communicate complex risk information to patients with low numeracy and visual literacy. However, the visualizations need to be carefully chosen to balance readability with ease of comprehension. In addition, due to patients' charged emotional state, an intuitive narrative structure that considers the patients' information need is critical to aid the patients' comprehension of their risk information.

Individualized risk prediction of outcomes for oral cavity cancer patients

BACKGROUND

Optimal management of oral cancer relies upon accurate and individualized risk prediction of relevant clinical outcomes. Individualized prognostic calculators have been developed to guide patient-physician communication and treatment-related decision-making. However it is critical to scrutinize their accuracy prior to integrating into clinical care.

AIM

To compare and evaluate oral cavity cancer prognostic calculators using an independent dataset.

METHODS

Five prognostic calculators incorporating patient and tumor characteristics were identified that evaluated five-year overall survival. A total of 505 patients with previously untreated oral cancer diagnosed between 2003 and 2014 were analyzed. Calculators were applied to each patient to generate individual predicted survival probabilities. Predictions were compared among prognostic tools and with observed outcomes using Kaplan-Meier plots, ROC curves and calibration plots.

RESULTS

Correlation between the five calculators varied from 0.59 to 0.86. There were considerable differences between individual predictions from pairs of calculators, with as many as 64% of patients having predictions that differed by more than 10%. Four of five calculators were well calibrated. For all calculators the predictions were associated with survival outcomes. The area under the ROC curve ranged from 0.65 to 0.71, with C-indices ranging from 0.63 to 0.67. An average of the 5 predictions had slightly better performance than any individual calculator.

CONCLUSION

Five prognostic calculators designed to predict individual outcomes of oral cancer differed significantly in their assessments of risk. Most were well calibrated and had modest discriminatory ability. Given the increasing importance of individualized risk prediction, more robust models are needed.

Refining prognosis after trans-arterial chemo-embolization for hepatocellular carcinoma

BACKGROUND & AIMS

To develop an individual prognostic calculator for patients with unresectable hepatocellular carcinoma (HCC) undergoing trans-arterial chemo-embolization (TACE).

METHODS

Data from two prospective databases, regarding 361 patients who received TACE as first-line therapy (2000-2012), were reviewed in order to refine available prognostic tools and to develop a continuous individual web-based prognostic calculator. Patients with neoplastic portal vein invasion were excluded from the analysis. The model was built following a bootstrap resampling procedure aimed at identifying prognostic predictors and by carrying out a 10-fold cross-validation for accuracy assessment by means of Harrell's c-statistic.

RESULTS

Number of tumours, serum albumin, serum total bilirubin, alpha-foetoprotein and maximum tumour size were selected as predictors of mortality following TACE with the bootstrap resampling technique. In the 10-fold cross-validation cohort, the model showed a Harrell's c-statistic of 0.649 (95% CI: 0.610-0.688), significantly higher than that of the Hepatoma Arterial-embolization Prognostic (HAP) score (0.589; 95% CI: 0.552-0.626; P = 0.001) and of the modified HAP-II score (0.611; 95% CI: 0.572-0.650; P = 0.005). Akaike's information criterion for the model was 2520; for the mHAP-II it was 2544 and for the HAP score it was 2554. A web-based calculator was developed for quick consultation at http://www.livercancer.eu/mhap3.html.

CONCLUSIONS

The proposed individual prognostic model can provide an accurate prognostic prediction for each patient with unresectable HCC following treatment with TACE without class stratification. The availability of an online calculator can help physicians in daily clinical practice.

The Surveillance, Epidemiology, and End Results Cancer Survival Calculator SEER*CSC: validation in a managed care setting

BACKGROUND

Nomograms for prostate and colorectal cancer are included in the Surveillance, Epidemiology, and End Results (SEER) Cancer Survival Calculator, under development by the National Cancer Institute. They are based on the National Cancer Institute's SEER data, coupled with Medicare data, to estimate the probabilities of surviving or dying from cancer or from other causes based on a set of patient and tumor characteristics. The nomograms provide estimates of survival that are specific to the characteristics of the tumor, age, race, gender, and the overall health of a patient. These nomograms have been internally validated using the SEER data. In this paper, we externally validate the nomograms using data from Kaiser Permanente Colorado.

METHODS

The SEER Cancer Survival Calculator was externally validated using time-dependent area under the Receiver Operating Characteristic curve statistics and calibration plots for retrospective cohorts of 1102 prostate cancer and 990 colorectal cancer patients from Kaiser Permanente Colorado.

RESULTS

The time-dependent area under the Receiver Operating Characteristic curve statistics were computed for one, three, five, seven, and 10 year(s) postdiagnosis for prostate and colorectal cancer and ranged from 0.77 to 0.89 for death from cancer and from 0.72 to 0.81 for death from other causes. The calibration plots indicated a very good fit of the model for death from cancer for colorectal cancer and for the higher risk group for prostate cancer. For the lower risk groups for prostate cancer (<10% chance of dying of prostate cancer in 10 years), the model predicted slightly worse prognosis than observed. Except for the lowest risk group for colorectal cancer, the models for death from other causes for both prostate and colorectal cancer predicted slightly worse prognosis than observed.

CONCLUSIONS

The results of the external validation indicated that the colorectal and prostate cancer nomograms are reliable tools for physicians and patients to use to obtain information on prognosis and assist in establishing priorities for both treatment of the cancer and other conditions, particularly when a patient is elderly and/or has significant comorbidities. The slightly better than predicted risk of death from other causes in a health maintenance organization (HMO) setting may be due to an overall healthier population and the integrated management of disease relative to the overall population (as represented by SEER).

From Foucault to Freire Through Facebook: Toward an Integrated Theory of mHealth

OBJECTIVE

To document the integration of social science theory in literature on mHealth (mobile health) and consider opportunities for integration of classic theory, health communication theory, and social networking to generate a relevant theory for mHealth program design.

METHOD

A secondary review of research syntheses and meta-analyses published between 2005 and 2014 related to mHealth, using the AMSTAR (A Measurement Tool to Assess Systematic Reviews) methodology for assessment of the quality of each review. High-quality articles from those reviews using a randomized controlled design and integrating social science theory in program design, implementation, or evaluation were reviewed. Results There were 1,749 articles among the 170 reviews with a high AMSTAR score (≥30). Only 13 were published from 2005 to 2014, used a randomized controlled design and made explicit mention of theory in any aspect of their mHealth program. All 13 included theoretical perspectives focused on psychological and/or psychosocial theories and constructs. Conclusions There is a very limited use of social science theory in mHealth despite demonstrated benefits in doing so. We propose an integrated theory of mHealth that incorporates classic theory, health communication theory, and social networking to guide development and evaluation of mHealth programs.

Patient-centered, evidence-based, and cost-conscious cancer care across the continuum: Translating the Institute of Medicine report into clinical practice

In 2013, the Institute of Medicine (IOM) concluded that cancer care in the United States is in crisis. Patients and their families are not receiving the information that they need to make informed decisions about their cancer care. Many patients do not have access to palliative care and too few are referred to hospice at the appropriate point in their disease trajectory. Simultaneously, there is a growing demand for cancer care with increases in new cancer diagnoses and the number of patients surviving cancer. Furthermore, there is a workforce shortage to care for this growing and elderly population. The IOM's report, Delivering High-Quality Cancer Care: Charting a New Course for a System in Crisis, outlined recommendations to improve the quality of cancer care. This article provides an overview of the IOM report and highlights the recommendations that are most relevant to practicing clinicians who care for patients with cancer across the continuum. The implementation of the recommendations in clinical practice will require better patient-clinician communication, improved care coordination, targeted clinician training, effective dissemination of evidence-based guidelines and strategies for eliminating waste, and continuous quality assessment and improvement efforts.

Providing clinicians and patients with actual prognosis: cancer in the context of competing causes of death

BACKGROUND

To isolate progress against cancer from changes in competing causes of death, population cancer registries have traditionally reported cancer prognosis (net measures). But clinicians and cancer patients generally want to understand actual prognosis (crude measures): the chance of surviving, dying from the specific cancer and from competing causes of death in a given time period.

OBJECTIVE

To compare cancer and actual prognosis in the United States for four leading cancers-lung, breast, prostate, and colon-by age, comorbidity, and cancer stage and to provide templates to help patients, clinicians, and researchers understand actual prognosis.

METHOD

Using population-based registry data from the Surveillance, Epidemiology, and End Results (SEER) Program, we calculated cancer prognosis (relative survival) and actual prognosis (five-year overall survival and the "crude" probability of dying from cancer and competing causes) for three important prognostic determinants (age, comorbidity [Charlson-score from 2012 SEER-Medicare linkage dataset] and cancer stage at diagnosis).

RESULT

For younger, healthier, and earlier stage cancer patients, cancer and actual prognosis estimates were quite similar. For older and sicker patients, these prognosis estimates differed substantially. For example, the five-year overall survival for an 85-year-old patient with colorectal cancer is 54% (cancer prognosis) versus 22% (actual prognosis)-the difference reflecting the patient's substantial chance of dying from competing causes. The corresponding five-year chances of dying from the patient's cancer are 46% versus 37%. Although age and comorbidity lowered actual prognosis, stage at diagnosis was the most powerful factor: The five-year chance of colon cancer death was 10% for localized stage and 83% for distant stage.

CONCLUSION

Both cancer and actual prognosis measures are important. Cancer registries should routinely report both cancer and actual prognosis to help clinicians and researchers understand the difference between these measures and what question they can and cannot answer. We encourage them to use formats like the ones presented in this paper to communicate them clearly.

Institute of medicine recommendations for improving the quality of cancer care: what do they mean for the general internist?

In order to evaluate and address the deficiencies in the U.S. cancer care system, particularly affecting the growing elderly population, the Institute of Medicine (IOM) convened a panel representing oncology providers, surgeons, primary care providers, researchers, policy makers and patients. The Committee concluded that cancer care is on the brink of crisis and issued recommendations targeting all stakeholders involved in cancer care. General internists play a critical role in the care of cancer patients, from the time of diagnosis, through treatment, survivorship and end of life care. We review the IOM recommendations, highlight those that are particularly relevant to the general internist, and outline clinical, research and educational opportunities where general internists should take an expanded role.

Risk prediction tools in surgical oncology

Healthcare has increasingly focused on patient engagement and shared decision-making. Decision aids can promote engagement and shared decision making by providing patients and their providers with care options and outcomes. This article discusses decision aids for surgical oncology patients. Topics include: short-term risk prediction following surgery, long-term risk prediction of survival and recurrence, the combination of short- and long-term risk prediction to help guide treatment choice, and decision aid usability, transparency, and accessibility.