Survival Prediction in Terminally Ill Cancer Patients by Clinical Estimates, Laboratory Tests, and Self-Rated Anxiety and Depression

LCD Benchmark: Long Clinical Document Benchmark on Mortality Prediction for Language Models

OBJECTIVE

The application of Natural Language Processing (NLP) in the clinical domain is important due to the rich unstructured information in clinical documents, which often remains inaccessible in structured data. When applying NLP methods to a certain domain, the role of benchmark datasets is crucial as benchmark datasets not only guide the selection of best-performing models but also enable the assessment of the reliability of the generated outputs. Despite the recent availability of language models (LMs) capable of longer context, benchmark datasets targeting long clinical document classification tasks are absent.

MATERIALS AND METHODS

To address this issue, we propose LCD benchmark, a benchmark for the task of predicting 30-day out-of-hospital mortality using discharge notes of MIMIC-IV and statewide death data. We evaluated this benchmark dataset using baseline models, from bag-of-words and CNN to instruction-tuned large language models. Additionally, we provide a comprehensive analysis of the model outputs, including manual review and visualization of model weights, to offer insights into their predictive capabilities and limitations.

RESULTS AND DISCUSSION

Baseline models showed 28.9% for best-performing supervised models and 32.2% for GPT-4 in F1-metrics. Notes in our dataset have a median word count of 1687. Our analysis of the model outputs showed that our dataset is challenging for both models and human experts, but the models can find meaningful signals from the text.

CONCLUSION

We expect our LCD benchmark to be a resource for the development of advanced supervised models, or prompting methods, tailored for clinical text. The benchmark dataset is available at https://github.com/Machine-Learning-for-Medical-Language/long-clinical-doc.

How accurate is clinical prognostication by oncologists during routine practice in a general hospital and can it be improved by a specific prognosis training programme: a prospective interventional study

OBJECTIVES

Oncologists need competence in clinical prognostication to deliver appropriate care to patients with cancer. Most studies on prognostication have been restricted to patients in palliative care settings. This paper investigates (1) the prognostic accuracy of physicians regarding a broad cohort of patients with cancer with a median life expectancy of >2 years and (2) whether a prognosis training can improve prognostication.

DESIGN

Prospective single-centre study comprising 3 phases, each lasting 1 month.

SETTING

Large teaching hospital, department of oncology and haematology, Germany.

PARTICIPANTS

18 physicians with a professional experience from entry level to 34 years. 736 patients with oncological and malignant haematological diseases.

INTERVENTIONS

Baseline prognostication abilities were recorded during an 'untrained' phase 1. As an intervention, a specific prognosis-training programme was implemented prior to phases 2 and 3. In phase 3, physicians had to provide additional estimates with the inclusion of electronic prognostic tools.

OUTCOME MEASURES

Prognostic estimates (PE) were collected using 'standard' surprise question (SQ), 'probabilistic' SQ (both for short-term prognostication up to 6 months) and clinician prediction of survival (CPS) (for long-term prognostication). Estimated prognoses were compared with observed survival. Phase 1 was compared with phases 2 and 3.

RESULTS

We included 2427 PE for SQ, 1506 for CPS and 800 for probabilistic SQ. Median OS was 2.5 years. SQ accuracy improved significantly (p<0.001) from 72.6% in phase 1 to 84.3% in phase 3. Probabilistic SQ in phase 3 showed 83.1% accuracy. CPS accuracy was 25.9% and could not be significantly improved. (Electronic) prognostic tools-used alone-performed significantly worse (p<0.0005) than physicians and-used by the clinicians-did not improve their performance.

CONCLUSION

A specific prognosis-training programme could improve short-term and intermediate-term prognostication. Improvement of long-term prognostication was not possible. Inexperienced residents as well as experienced oncologists benefited from training.

Thirty-day mortality as a metric for palliative radiotherapy in pediatric patients

PURPOSE OF REVIEW

Thirty-day mortality (30DM) is an emerging consideration for determining whether terminally ill adult patients may benefit from palliative radiotherapy (RT). However, the efficacy and ethics of delivering palliative RT at the end of life (EOL) in children are seldom discussed and not well-established.

RECENT FINDINGS

Palliative RT is perhaps underutilized among patients ≤21 years old with rates as low as 11%. While effective when delivered early, clinical benefit decreases when administered within the last 30 days of life. Pediatric 30DM rates vary widely between institutions (0.7-30%), highlighting the need for standardized practices. Accurate prognosis estimation remains challenging and prognostic models specific to palliative pediatric patients are limited. Discordance between provider and patient/parent perceptions of prognosis further complicates decision-making.

SUMMARY

RT offers effective symptom control in pediatric patients when administered early. However, delivering RT within the last 30 days of life may provide limited clinical benefit and hinder optimal EOL planning and care. Early referral for palliative RT, preferably with fewer fractions (five or fewer), along with multidisciplinary supportive care, optimizes the likelihood of maintaining patients' quality of life. Prognosis estimation remains difficult, and improving patient and family understanding is crucial. Further research is needed to refine prognostic models and enhance patient-centered care.

Usefulness of the Palliative Prognostic Index in Predicting Prognosis when Considering the Transition from Hospital to Home Care in Patients with Terminal Stage Cancer

BACKGROUND

No accurate prognostic tool is available for patients with cancer who spend their final days at home. In this study, we examined whether performance status (PS) and the palliative prognostic index (PPI), a well-known prognostic tool in palliative care units, could be used to predict prognosis in the home care setting at the time of intervention by home physicians.

SUBJECTS AND METHODS

Using medical records, we conducted a retrospective analysis of 132 patients who were referred to the Home Clinic Naginoki for home care for terminal stages of carcinoma in situ. Based on the status at the time of the first visit, the PPI-Low group was defined as those scoring six or below and the PPI-High group as those scoring greater than six.

RESULTS

The PPI-high group had a significantly poorer prognosis within 21 days than the PPI-low group (21-day-OS; Low 71.4% vs. High 13.2%; p<0.001). The Eastern Cooperative Oncology Group (ECOG) PS alone predicted better prognosis in the group with PS of one or two (21-day survival 90.1%), and the PPI score further significantly stratified the prognosis for patients with PS three or four, with a trend toward poor prognosis (p ≤ 0.005).

CONCLUSION

ECOG PS 1 or 2 has a favorable prognosis and that using PPI in ECOG PS 3 or 4 leads to a more accurate prognosis prediction. PPI evaluated during the hospital-based treatment of patients with terminal cancer can also be used to predict prognosis if the patient is transitioned to a home care environment.

[ESPEN practical guideline: Home enteral nutrition]

Introduction

This ESPEN practical guideline will inform physicians, nurses, dieticians, pharmacists, caregivers and other home enteral nutrition (HEN) providers in a concise way about the indications and contraindications for HEN, as well as its implementation and monitoring. This guideline will also inform interested patients requiring HEN. Home parenteral nutrition is not included but will be addressed in a separate ESPEN guideline. The guideline is based on the ESPEN scientific guideline published before, which consists of 61 recommendations that have been reproduced and renumbered, along with the associated commentaries that have been shorted compared to the scientific guideline. Evidence grades and consensus levels are indicated. The guideline was commissioned and financially supported by ESPEN and the members of the guideline group were selected by ESPEN.

Could palliative sedation be seen as unnamed euthanasia?: a survey among healthcare professionals in oncology

BACKGROUND

In 2016 a French law created a new right for end-of-life patients: deep and continuous sedation maintained until death, with discontinuation of all treatments sustaining life such as artificial nutrition and hydration. It was totally unprecedented that nutrition and hydration were explicitly defined in France as sustaining life treatments, and remains a specificity of this law. End- of-life practices raise ethical and practical issues, especially in Europe actually. We aimed to know how oncology professionals deal with the law, their opinion and experience and their perception.

METHODS

Online mono-centric survey with closed-ended and open-ended questions in a Cancer Comprehensive Centre was elaborated. It was built during workshops of the ethics committee of the Institute, whose president is an oncologist with a doctoral degree in medical ethics. 58 oncologists and 121 nurses-all professionals of oncological departments -, received it, three times, as mail, with an information letter.

RESULTS

63/ 179 professionals answered the questionnaire (35%). Conducting end-of-life discussions and advanced care planning were reported by 46/63 professionals. In the last three months, 18 doctors and 7 nurses faced a request for a deep and continuous sedation maintained until death, in response to physical or existential refractory suffering. Artificial nutrition and even more hydration were not uniformly considered as treatment. Evaluation of the prognosis, crucial to decide a deep and continuous sedation maintained until death, appears to be very difficult and various, between hours and few weeks. Half of respondents were concerned that this practice could lead to or hide euthanasia practices, whereas for the other half, this new law formalised practices necessary for the quality of palliative care at the end-of-life.

CONCLUSION

Most respondents support the implementation of deep and continuous sedation maintained until death in routine end-of-life care. Nevertheless, difficulty to stop hydration, confusion with euthanasia practices, ethical debates it provokes and the risk of misunderstanding within teams and with families are significant. This is certainly shared by other teams. This could lead to a multi-centric survey and if confirmed might be reported to the legislator.

Early death after palliative radiation treatment: 30-, 35- and 40-day mortality data and statistically robust predictors

BACKGROUND

This study analyzed mortality after radiotherapy for bone metastases (287 courses). Endpoints such as treatment in the last month of life and death within 30, 35 and 40 days from start of radiotherapy were evaluated.

METHODS

Different baseline parameters including but not limited to blood test results and patterns of metastases were assessed for association with early death. After univariate analyses, multi-nominal logistic regression was employed.

RESULTS

Of 287 treatment courses, 42 (15%) took place in the last month of life. Mortality from start of radiotherapy was 13% (30-day), 15% (35-day) and 18% (40-day), respectively. We identified three significant predictors of 30-day mortality (performance status (≤ 50, 60-70, 80-100), weight loss of at least 10% within 6 months (yes/no), pleural effusion (present/absent)) and employed these to construct a predictive model with 5 strata and mortality rates of 0-75%. All predictors of 30-day mortality were also associated with both, 35- and 40-day mortality.

CONCLUSION

Early death was not limited to the first 30 days after start of radiotherapy. For different cut-off points, similar predictive factors emerged. A model based on three robust predictors was developed.

Radiotherapy for Lung Metastases: Conventional to Stereotactic Body Radiation Therapy

The lung parenchyma and adjacent tissues are one of the most common sites of metastatic disease. Traditionally, the approach to treatment of a patient with lung metastases has been with systemic therapy, with radiotherapy being reserved for palliative management of symptomatic disease. The concept of oligo metastatic disease has paved the way for more radical treatment options, administered either alone or as local consolidative therapy in addition to systemic treatment. The modern-day management of lung metastases is guided by a number of factors, including the number of lung metastases, extra-thoracic disease status, overall performance status, and life expectancy, which all help determine the goals of care. Stereotactic body radiotherapy (SBRT) has emerged as a safe and effective method in locally controlling lung metastases, in the oligo metastatic or oligo-recurrent setting. This article outlines the role of radiotherapy in multimodality management of lung metastases.

Clinical outcome and prognostic factors for Asian patients in Phase I clinical trials

BACKGROUND

Patient selection is key in Phase I studies, and prognosis can be difficult to estimate in heavily pre-treated patients. Previous prognostic models like the Royal Marsden Hospital (RMH) score or using the neutrophil-lymphocyte ratio (NLR) have not been validated in current novel therapies nor in the Asian Phase I population.

METHODS

We conducted a retrospective review of 414 patients with solid tumours participating in Phase I studies at our centre between October 2013 and December 2020.

RESULTS

The RMH model showed poorer prognosis with increasing scores [RMH score 1, HR 1.28 (95% CI: 0.96-1.70); RMH score 2, HR 2.27 (95% CI: 1.62-3.17); RMH score 3, HR 4.14 (95% CI: 2.62-6.53)]. NLR did not improve the AUC of the model. Poorer ECOG status (ECOG 1 vs. 0: HR = 1.59 (95% CI = 1.24-2.04), P < 0.001) and primary tumour site (GI vs. breast cancer: HR = 3.06, 95% CI = 2.16-4.35, P < 0.001) were prognostic.

CONCLUSIONS

We developed a NCIS prognostic score with excellent prognostic ability for both short-term and longer-term survival (iAUC: 0.71 [95% CI 0.65-0.76]), and validated the RMH model in the largest Asian study to date.

The link between psychological distress and survival in solid tumor patients: A systematic review

PURPOSE

Research has demonstrated that solid tumor patients experience high levels of psychological distress at the time of diagnosis. While distress has been associated with many adverse clinical outcomes, little is known about how this symptom may influence the disease trajectory for cancer patients, affecting outcomes such as progression, recurrence, and survival. The purpose of this systematic review was to explore the literature linking distress with survival in solid tumor patients, which may guide future work exploring clinical outcomes as a function of distress.

METHODS

A systematic search of PubMed, Embase, and Web of Science was performed using PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines with predefined eligibility criteria. Thirteen studies met the inclusion criteria and were selected for review.

RESULTS

Findings from this review demonstrated a weak-to-moderate relationship between cancer patients' experience of distress and overall survival, with most included studies (11/13) finding at least one predictive analysis to be significant when controlling for confounders. However, significant heterogeneity in the literature, particularly with study sample characteristics and varying methodologies, made direct comparisons across studies challenging.

CONCLUSION

Findings from this review suggest that psychological distress may have an impact on disease-related outcomes, including (but not limited to) survival. Future work should consider performing disease-specific analyses controlling for key prognostic factors to better understand the nuanced relationship between distress and clinical outcomes, which may allow further understanding of the biological underpinnings of this relationship and enable the development of targeted interventions for improving distress.

Prognostic factors in patients in the terminal phase of haematological malignancies who are receiving home medical care

Although there are many prognostic models for patients in the terminal phase of solid tumours, a reliable prognostic scoring system in patients in the terminal phase of haematological malignancies (HM) has not been established. We retrospectively evaluated 180 patients in the terminal phase of HM who were receiving home medical care (HMC). Multivariate analyses revealed that clinician's estimate, consciousness, loss of appetite, dyspnoea, neutrophil count, lymphocyte count, and lactate dehydrogenase were associated with overall survival (OS). Based on this result, we developed a novel prognostic scoring system, the Japan palliative haematological oncology prognostic estimates, in which four risk groups were shown to clearly differ in survival (p < 0.001): a low-risk group (n = 41, median OS of 434 days), an intermediate-low-risk group (n = 80, median OS of 112 days), an intermediate-high-risk group (n = 38, median OS of 31.5 days), and a high-risk group (n = 21, median OS of 10 days). This is the first investigation of prognostic factors that influence the OS of patients in the terminal phase of HM who are receiving HMC. Providing patients with reliable information about their prognosis is important for them to consider how to spend their remaining life.

Factors Influencing Physician Prognosis: A Scoping Review

Introduction. Prognosis is an essential component of informed consent for medical decision making. Research shows that physicians display discrepancies in their prognostication, leading to variable, inaccurate, optimistic, or pessimistic prognosis. Factors driving these discrepancies and the supporting evidence have not been reviewed systematically. Methods. We undertook a scoping review to explore the literature on the factors leading to discrepancies in medical prognosis. We searched Medline (Ovid) and Embase (Ovid) databases for peer-reviewed articles from 1970 to 2017. We included articles that discussed prognosis variation or discrepancy and where factors influencing prognosis were evaluated. We extracted data outlining the participants, methodology, and prognosis discrepancy information and measured factors influencing prognosis. Results. Of 4,723 articles, 73 were included in the final analysis. There was significant variability in research methodologies. Most articles showed that physicians were pessimistic regarding patient outcomes, particularly in early trainees and acute care specialties. Accuracy rates were similar across all time periods. Factors influencing prognosis were clustered in 4 categories: patient-related factors (such as age, gender, race, diagnosis), physician-related factors (such as age, race, gender, specialty, training and experience, attitudes and values), clinical situation-related factors (such as physician-patient relationship, patient location, and clinical context), and environmental-related factors (such as country or hospital size). Discussion. Obtaining accurate prognostic information is one of the highest priorities for seriously ill patients. The literature shows trends toward pessimism, especially in early trainees and acute care specialties. While some factors may prove difficult to change, the physician's personality and psychology influence prognosis accuracy and could be tackled using debiasing strategies. Exposure to long-term patient outcomes and a multidisciplinary practice setting are environmental debiasing strategies that may warrant further research.

Highlights

Literature on discrepancies in physician's prognostication is heterogeneous and sparse.Literature shows that physicians are mostly pessimistic regarding patient outcomes.Literature shows that a physician's personality and psychology influence prognostic accuracy and could be improved with evidence-based debiasing strategies.Medical specialty strongly influences prognosis, with specialties exposed to acutely ill patients being more pessimistic, whereas specialties following patients longitudinally being more optimistic.Physicians early in their training were more pessimist than more experienced physicians.

The Prognostic Utility of a Geriatric Assessment for Patients with Pancreatic Cancer Receiving Gemcitabine-based Chemotherapy: A Prospective Observational Study

Objective To evaluate the effect of a pretreatment geriatric assessment on the clinical outcomes in older patients with unresectable or recurrent pancreatic ductal adenocarcinoma (PDAC) scheduled to receive gemcitabine (GEM)/GEM+nab-paclitaxel (GnP). Patients Older patients with unresectable PDAC scheduled to receive GEM/GnP who visited Kyorin University Hospital and cooperating institutions were enrolled and followed from April 2015 to March 2020. The maximum observation period was two years. All patients underwent a cancer-specific geriatric assessment (CSGA) and optional geriatric assessment (GA) before treatment initiation and two months after the start of treatment. The patients' background characteristics, tumor progression, tumor site, and regimen (GEM/GnP) were examined in a Cox proportional hazards model. The relationship between the overall survival (OS) and GA score was also determined. Eligible patients (age ≥70 years old with histopathologically confirmed unresectable or recurrent PDAC) were scheduled to receive first-line chemotherapy. Results The performance status (PS) and activities of daily living (ADL)/instrumental ADL (IADL) scores at baseline correlated with the OS. Furthermore, even in cases with normal baseline values, lower Frontal Assessment Battery scores and higher Geriatric Depression Scale-Short Form scores after treatment initiation were significantly correlated with OS. Conclusion The baseline PS, ADL, and IADL may be prognostic factors in older PDAC patients. In addition, a normal frontal lobe function and depression scores prior to treatment initiation that rapidly worsened during treatment were independently associated with a reduced OS. Selecting appropriate interventions and improving the therapeutic environment may prolong the OS in such patients.

Shock Index Is a Validated Prediction Tool for the Short-Term Survival of Advanced Cancer Patients Presenting to the Emergency Department

Advanced cancer patients who are not expected to survive past the short term can benefit from early initiation of palliative care in the emergency department (ED). This discussion, however, requires accurate prognostication of their short-term survival. We previously found in our retrospective study that shock index (SI) is an ideal risk stratification tool in predicting the 60-day mortality risk of advanced cancer patients presenting to the ED. This study is a follow-up prospective validation study conducted from January 2019 to April 2021. A total of 410 advanced cancer patients who presented to the ED of a medical centre and could be followed-up feasibly were recruited. Univariate and multivariable logistic regression analyses were performed with receiver operator calibrating (ROC) curve analysis. Non-survivors had significantly lower body temperatures, higher pulse rates, higher respiratory rates, lower blood pressures, and higher SI. Each 0.1 increment of SI increased the odds of 60-day mortality by 1.591. Area under ROC curve was 0.7819. At optimal cut-off of 0.94, SI had 66.10% accuracy. These results were similar to our previous study, thus validating the use of SI in predicting the 60-day mortality of advanced cancer patients presenting to the ED. Identified patients may be offered palliative care.

Prevalence and predictors for 72-h mortality after transfer to acute palliative care unit

Purpose

Accurate prediction of survival is important to facilitate clinical decision-making and improve quality of care at the end of life. While it is well documented that survival prediction poses a challenge for treating physicians, the need for clinically valuable predictive factors has not been met. This study aims to quantify the prevalence of patient transfer 72 h before death onto the acute palliative care unit in a tertiary care center in Switzerland, and to identify factors predictive of 72-h mortality.

Methods

All patients hospitalized between January and December 2020 on the acute palliative care unit of the Competence Center Palliative Care of the Department of Radiation Oncology at the University Hospital Zurich were assessed. Variables were retrieved from the electronic medical records. Univariable and multivariable logistic regressions were used to identify predictors of mortality.

Results

A total of 398 patients were screened, of which 188 were assessed. Every fifth patient spent less than 72 h on the acute palliative care unit before death. In multivariable logistic regression analysis, predictors for 72-h mortality after transfer were no prior palliative care consult (p = 0.011), no advance care directive (p = 0.044), lower performance status (p = 0.035), lower self-care index (p = 0.003), and lower blood albumin level (p = 0.026).

Conclusion

Late transfer to the acute palliative care unit is not uncommon, which can cause additional distress to patients and caretakers. Though clinically practical short-term survival predictors remain largely unidentified, early integration of palliative care should be practiced more regularly in patients with life-limiting illness.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00520-022-07075-6.

Multidisciplinary Treatment of Non-Spine Bone Metastases: Results of a Modified Delphi Consensus Process

•

Evidence is emerging for new paradigms in the management of non-spine bone metastases.

•

Consensus was feasible amongst physicians in both academic and community-based practice settings.

•

Topics deemed of highest importance for consensus included referral for surgical stabilization and approach to peri-operative radiation, preferred radiation fractionation and appropriate use of stereotactic techniques, and clinical scenarios classified as potentially “complex” warranting multidisciplinary discussion.

Purpose

Methods and Materials

Results

Conclusions

Prediction of post-radiotherapy survival for bone metastases: a comparison of the 3-variable number of risk factors model with the new Katagiri scoring system

We investigated patient survival after palliative radiotherapy for bone metastases while comparing the prognostic accuracies of the 3-variable number of risk factors (NRF) model and the new Katagiri scoring system (Katagiri score). Overall, 485 patients who received radiotherapy for bone metastases were grouped as per the NRF model (groups I, II and III) and Katagiri score (low-, intermediate- and high-risk). Survival was compared using the log-rank or log-rank trend test. Independent prognostic factors were identified using multivariate Cox regression analyses (MCRA). MCRA and receiver operating characteristic (ROC) curves were used to compare both models' accuracy. For the 376 evaluable patients, the overall survival (OS) rates decreased significantly in the higher-tier groups of both models (P < 0.001). All evaluated factors except 'previous chemotherapy status' differed significantly between groups. Both models exhibited independent predictive power (P < 0.001). Per NRF model, hazard ratios (HRs) were 1.44 (P = 0.099) and 2.944 (P < 0.001), respectively, for groups II and III, relative to group I. Per Katagiri score, HRs for intermediate- and high-risk groups were 4.02 (P < 0.001) and 7.09 (P < 0.001), respectively, relative to the low-risk group. Areas under the curve (AUC) for predicting 6-, 18- and 24-month mortality were significantly higher when using the Katagiri score (P = 0.036, 0.039 and 0.022). Both models predict survival. Prognostic accuracy of the Katagiri score is superior, especially in patients with long-term survival potential; however, in patients with short prognosis, no difference occurred between both models; simplicity and patient burden should also be considered.

Association of Prognostic Understanding With Health Care Use Among Older Adults With Advanced Cancer: A Secondary Analysis of a Cluster Randomized Clinical Trial

IMPORTANCE

A poor prognostic understanding regarding curability is associated with lower odds of hospice use among patients with cancer. However, the association between poor prognostic understanding or prognostic discordance and health care use among older adults with advanced incurable cancers is not well characterized.

OBJECTIVE

To evaluate the association of poor prognostic understanding and patient-oncologist prognostic discordance with hospitalization and hospice use among older adults with advanced cancers.

DESIGN, SETTING, AND PARTICIPANTS

This was a post hoc secondary analysis of a cluster randomized clinical trial that recruited patients from October 29, 2014, to April 28, 2017. Data were collected from community oncology practices affiliated with the University of Rochester Cancer Center National Cancer Institute Community Oncology Research Program. The parent trial enrolled 541 patients who were aged 70 years or older and were receiving or considering any line of cancer treatment for incurable solid tumors or lymphomas; the patients' oncologists and caregivers (if available) were also enrolled. Patients were followed up for at least 1 year. Data were analyzed from January 3 to 16, 2021.

MAIN OUTCOMES AND MEASURES

At enrollment, patients and oncologists were asked about their beliefs regarding cancer curability (100%, >50%, 50%, <50%, and 0%; answers other than 0% reflected poor prognostic understanding) and life expectancy (≤6 months, 7-12 months, 1-2 years, 2-5 years, and >5 years; answers of >5 years reflected poor prognostic understanding). Any difference between oncologist and patient in response options was considered discordant. Outcomes were any hospitalization and hospice use at 6 months captured by the clinical research associates.

RESULTS

Among the 541 patients, the mean (SD) age was 76.6 (5.2) years, 264 of 540 (49%) were female, and 486 of 540 (90%) were White. Poor prognostic understanding regarding curability was reported for 59% (206 of 348) of patients, and poor prognostic understanding regarding life expectancy estimates was reported for 41% (205 of 496) of patients. Approximately 60% (202 of 336) of patient-oncologist dyads were discordant regarding curability, and 72% (356 of 492) of patient-oncologist dyads were discordant regarding life expectancy estimates. Poor prognostic understanding regarding life expectancy estimates was associated with lower odds of hospice use (adjusted odds ratio, 0.30; 95% CI, 0.16-0.59). Discordance regarding life expectancy estimates was associated with greater odds of hospitalization (adjusted odds ratio, 1.64; 95% CI, 1.01-2.66).

CONCLUSIONS AND RELEVANCE

This study highlights different constructs of prognostic understanding and the need to better understand the association between prognostic understanding and health care use among older adult patients with advanced cancer.

TRIAL REGISTRATION

ClinicalTrials.gov Identifier: NCT02107443.

Patient, Caregiver, and Oncologist Predictions of Quality of Life in Advanced Cancer: Accuracy and Associations with End-of-Life Care and Caregiver Bereavement

OBJECTIVE

Informed treatment decision-making necessitates accurate prognostication,including predictions about quality of life. We examined whether oncologists, patients with advanced cancer, and caregivers accurately predict patients' future quality of life and whether these predictions are prospectively associated with end-of-life care and bereavement.

METHODS

We secondary analyses of clinical trial data. Patients with advanced cancer (n=156), caregivers (n=156), and oncologists (n=38) predicted patient quality of life 3 months into the future. Patients subsequently rated their quality of life 3 months later. Medical record data documented chemotherapy and emergency department (ED)/inpatient visits in the 30 days before death (n=79 decedents). Caregivers self-reported on depression, anxiety, grief, purpose, 21 and regret 7-months post-mortem. In mixed-effects models, patient, caregiver, and oncologist quality-of-life predictions at study entry were used to predict end-of-life care and caregiver outcomes, controlling for patients' quality of life at 3-month follow-up, demographic and clinical characteristics, and nesting within oncologists.

RESULTS

Caregivers (P<.0001) and oncologists (P=.001) predicted lower quality of life than what patients actually experienced. Among decedents, 24.0% received chemotherapy and 54.5% had an ED/inpatient visit. When caregivers' predictions were more negative, patients were less likely to receive chemotherapy (P=.028) or have an ED/inpatient visit (P=.033), and caregivers reported worse depression (P=.002), anxiety (P=.019), and grief (P=.028) and less purpose in life (P<.001) 7-months post-mortem.

CONCLUSION

When caregivers have more negative expectations about patients' quality of life, patients receive less intensive end-of-life care, and caregivers report worse bereavement This article is protected by copyright. All rights reserved.

Predictors of Mortality in Patients with Advanced Cancer-A Systematic Review and Meta-Analysis

To timely initiate advance care planning in patients with advanced cancer, physicians should identify patients with limited life expectancy. We aimed to identify predictors of mortality. To identify the relevant literature, we searched Embase, MEDLINE, Cochrane Central, Web of Science, and PubMed databases between January 2000-April 2020. Identified studies were assessed on risk-of-bias with a modified QUIPS tool. The main outcomes were predictors and prediction models of mortality within a period of 3-24 months. We included predictors that were studied in ≥2 cancer types in a meta-analysis using a fixed or random-effects model and summarized the discriminative ability of models. We included 68 studies (ranging from 42 to 66,112 patients), of which 24 were low risk-of-bias, and 39 were included in the meta-analysis. Using a fixed-effects model, the predictors of mortality were: the surprise question, performance status, cognitive impairment, (sub)cutaneous metastases, body mass index, comorbidity, serum albumin, and hemoglobin. Using a random-effects model, predictors were: disease stage IV (hazard ratio [HR] 7.58; 95% confidence interval [CI] 4.00-14.36), lung cancer (HR 2.51; 95% CI 1.24-5.06), ECOG performance status 1+ (HR 2.03; 95% CI 1.44-2.86) and 2+ (HR 4.06; 95% CI 2.36-6.98), age (HR 1.20; 95% CI 1.05-1.38), male sex (HR 1.24; 95% CI 1.14-1.36), and Charlson comorbidity score 3+ (HR 1.60; 95% CI 1.11-2.32). Thirteen studies reported on prediction models consisting of different sets of predictors with mostly moderate discriminative ability. To conclude, we identified reasonably accurate non-tumor specific predictors of mortality. Those predictors could guide in developing a more accurate prediction model and in selecting patients for advance care planning.

ESPEN practical guideline: Home enteral nutrition

This ESPEN practical guideline will inform physicians, nurses, dieticians, pharmacists, caregivers and other home enteral nutrition (HEN) providers in a concise way about the indications and contraindications for HEN, as well as its implementation and monitoring. This guideline will also inform interested patients requiring HEN. Home parenteral nutrition is not included but will be addressed in a separate ESPEN guideline. The guideline is based on the ESPEN scientific guideline published before, which consists of 61 recommendations that have been reproduced and renumbered, along with the associated commentaries that have been shorted compared to the scientific guideline. Evidence grades and consensus levels are indicated. The guideline was commissioned and financially supported by ESPEN and the members of the guideline group were selected by ESPEN.

Prospective Study Comparing Clinicians' and Cancer Patients' Estimates of Risk of Relapse and Toxicity with Adjuvant Chemotherapy

The aim of this study was to compare physicians' and patients' estimates of risk of relapse and toxicity. A prospective, cross-sectional, multicenter study including 735 patients with cancer and 29 oncologists. Physicians' appraisals of risk of relapse with and without chemotherapy (27.5% and 43.1%) and risk of severe toxicity (12.2%) were more realistic than those of patients (34.6%, 78.5%, and 57.4%, respectively). The greater the risk of recurrence and risk of toxicity estimated, the less physicians expressed satisfaction with SDM. Estimations of risk of relapse and toxicity are important in diagnostic and therapeutic decision-making and can help patients face their situation.

Early and Midtreatment Mortality in Palliative Radiotherapy: Emphasizing Patient Selection in High-Quality End-of-Life Care

BACKGROUND

Palliative radiotherapy (RT) is effective, but some patients die during treatment or too soon afterward to experience benefit. This study investigates end-of-life RT patterns to inform shared decision-making and facilitate treatment consistent with palliative goals.

MATERIALS AND METHODS

All patients who died ≤6 months after initiating palliative RT at an academic cancer center between 2015 and 2018 were identified. Associations with time-to-death, early mortality (≤30 days), and midtreatment mortality were analyzed.

RESULTS

In total, 1,620 patients died ≤6 months from palliative RT initiation, including 574 (34%) deaths at ≤30 days and 222 (14%) midtreatment. Median survival was 43 days from RT start (95% CI, 41-45) and varied by site (P<.001), ranging from 36 (head and neck) to 53 days (dermal/soft tissue). On multivariable analysis, earlier time-to-death was associated with osseous (hazard ratio [HR], 1.33; P<.001) and head and neck (HR, 1.45; P<.001) sites, multiple RT courses ≤6 months (HR, 1.65; P<.001), and multisite treatments (HR, 1.40; P=.008), whereas stereotactic technique (HR, 0.77; P<.001) and more recent treatment year (HR, 0.82; P<.001) were associated with longer survival. No difference in time to death was noted among patients prescribed conventional RT in 1 to 10 versus >10 fractions (median, 40 vs 47 days; P=.272), although the latter entailed longer courses. The 30-day mortality group included 335 (58%) inpatients, who were 27% more likely to die midtreatment (P=.031). On multivariable analysis, midtreatment mortality among these inpatients was associated with thoracic (odds ratio [OR], 2.95; P=.002) and central nervous system (CNS; OR, 2.44; P=.002) indications, >5-fraction courses (OR, 3.27; P<.001), and performance status of 3 to 4 (OR, 1.63; P=.050). Conversely, palliative/supportive care consultation was associated with decreased midtreatment mortality (OR, 0.60; P=.045).

CONCLUSIONS

Earlier referrals and hypofractionated courses (≤5-10 treatments) should be routinely considered for palliative RT indications, given the short life expectancies of patients at this stage in their disease course. Providers should exercise caution for emergent thoracic and CNS indications among inpatients with poor prognoses due to high midtreatment mortality.

Associations between the gut microbiome and fatigue in cancer patients

Fatigue is the most prevalent symptom of cancer and its treatments. Changes in the intestinal microbiome have been identified in chronic fatigue syndrome and other neuropsychiatric disorders, and cancer patients. However, the association between intestinal microbiome and fatigue in patients with advanced cancers has not been evaluated. Understanding the connection between intestinal microbiome and fatigue will provide interventional and therapeutic opportunities to manipulate the microbiome to improve fatigue and other patients’ reported outcomes. In this project, we aimed to identify associations between microbiome composition and fatigue in advanced cancer patients. In this cross-sectional observational study at a tertiary cancer care center, we included 88 patients with advanced, metastatic, unresectable cancers who were in a washout period from chemotherapy. We measured fatigue using the MD Anderson Symptom Inventory—Immunotherapy fatigue score, and used 16srRNA to analyze intestinal microbiome. Using correlation analysis we found that Eubacterium hallii was negatively associated with fatigue severity scores (r = − 0.30, p = 0.005), whereas Cosenzaea was positively associated with fatigue scores (r = 0.33, p = 0.0002). We identified microbial species that exhibit distinct composition between high-fatigued and low-fatigued cancer patients. Further studies are warranted to investigate whether modulating the microbiome reduces cancer patients’ fatigue severity and improves their quality of life.

Prognostic value of circulating markers of neutrophil activation, neutrophil extracellular traps, coagulation and fibrinolysis in patients with terminal cancer

Predicting survival accurately in patients with advanced cancer is important in guiding interventions and planning future care. Objective tools are therefore needed. Blood biomarkers are appealing due to their rapid measurement and objective nature. Thrombosis is a common complication in cancer. Recent data indicate that tumor-induced neutrophil extracellular traps (NETs) are pro-thrombotic. We therefore performed a comprehensive investigation of circulating markers of neutrophil activation, NET formation, coagulation and fibrinolysis in 106 patients with terminal cancer. We found that neutrophil activation and NET markers were prognostic in terminal cancer patients. Interestingly, markers of coagulation and fibrinolysis did not have a prognostic value in this patient group, and there were weak or no correlations between these markers and markers of neutrophil activation and NETs. This suggest that NETs are linked to a poor prognosis through pathways independent of coagulation. Additional studies are needed to determine the utility of circulating neutrophil activation and NET markers, alone or in concert with established clinical parameters, as objective and reliable prognostic tools in advanced cancer.

Impact of Prognostic Nutritional Index on Terminal Cancer Patients

Background:

In terminal cancer patients (TCPs), one of the most important things is to define the survival to help the main responsible physicians, patients, and main caregivers make decisions, set goals, and work across the end-of-life strategies.

Patients and Methods:

We retrospectively reviewed the medical files of TCPs, who died during September 2011 and December 2017, to recognize the correlation between prognostic nutritional indices (PNIs) and survival in those subtypes of patients. The receiver operating characteristic (ROC) curve was used to identify the cutoff value of PNI.

Results:

A total of 858 TCPs were eligible and included, the median age was 62 years (range: 18–107). The most common primary cancer sites were colorectal cancer in 151 patients (17.6%), hepatobiliary in 129 (15%), lung cancer in 115 (13.4%), breast cancer in 114 (13.3%), and genitourinary in 80 (9.3%). The mean value of PNI for all cancer types was 32.9 ± 6.7. The values showed different levels across cancer types. For patients who lived >2 weeks, PNI was 36.7 compared with that who died within 2 weeks was 29.3, which was a statistically significant (P < 0.001). By the ROC curve, the cutoff value of PNI was 32.3 and area under the curve was 0.888. The sensitivity, specificity, positive predictive value, and negative predictive value were 91.28% (95% confidence interval [CI]: 88.2–93.8), 71.09% (95% CI: 66.5–75.4), 76.5% (95% CI: 73.7–79.2), and 88.8% (95% CI: 85.3–91.5), respectively.

Conclusion:

The PNI is an easy and an applicable biomarker to estimate life expectancy in TCPs.

Confounding factors in exposure-response analyses and mitigation strategies for monoclonal antibodies in oncology

Dose selection and optimization is an important topic in drug development to maximize treatment benefits for all patients. While exposure-response (E-R) analysis is a useful method to inform dose-selection strategy, in oncology, special considerations for prognostic factors are needed due to their potential to confound the E-R analysis for monoclonal antibodies. The current review focuses on 3 different approaches to mitigate the confounding effects for monoclonal antibodies in oncology: (i) Cox-proportional hazards modelling and case-matching; (ii) tumour growth inhibition-overall survival modelling; and (iii) multiple dose level study design. In the presence of confounding effects, studying multiple dose levels may be required to reveal the true E-R relationship. However, it is impractical for pivotal trials in oncology drug development programmes. Therefore, the strengths and weaknesses of the other 2 approaches are considered, and the favourable utility of tumour growth inhibition-overall survival modelling to address confounding in E-R analyses is described. In the broader scope of oncology drug development, this review discusses the downfall of the current emphasis on E-R analyses using data from single dose level trials and proposes that development programmes be designed to study more dose levels in earlier trials.

Physician-Predicted Prognosis and Palliative Radiotherapy Treatment Utilization at the End of Life: An Audit of a Large Cancer Center Network

CONTEXT

At our institution, clinical pathways capture physicians' prognostication of patients being evaluated for palliative radiotherapy. We hypothesize a low utilization rate of long-course radiotherapy (LCRT) and stereotactic ablative radiotherapy (SAbR) among patients seen at the end of life, especially those with physician-predicted poor prognosis.

OBJECTIVE

To analyze utilization rates and predictors of LCRT and SAbR at the end of life.

METHODS

A retrospective review was conducted on patients who were evaluated for palliative radiotherapy between January 2017 and August 2019 and died within 90 days of consultation. Binary logistic regression was used to identify predictors for utilization of LCRT (≥10 fractions) and SAbR.

RESULTS

A total of 1608 patients were identified, of which 1038 patients (64.6%) were predicted to die within a year. Six hundred ninety-three patients (66.8%) out of 1038 were prescribed LCRT or SAbR. On a multivariate analysis, patients were less likely to be prescribed LCRT if treated at an academic site (odds ratio [OR], 0.30; 95% confidence interval [CI], 0.23-0.39; P < 0.01) and treated for bone metastases (OR, 0.08; 95% CI, 0.05-0.11; P < 0.01) or other nonbrain/nonbone metastases (OR, 0.19; 95% CI, 0.13-0.30; P < 0.01). SAbR was less likely to be prescribed among patients predicted to die within a year (OR, 0.09; 95% CI, 0.06-0.16; P < 0.01), treated for bone metastases (OR, 0.13; 95% CI, 0.07-0.22; P < 0.01), with poor performance status (OR, 0.51; 95% CI, 0.31-0.85; P = 0.01), and with a breast primary (OR, 0.35; 95% CI, 0.15-0.82; P = 0.02).

CONCLUSION

Although most patients were predicted to have a limited prognosis, LCRT and SAbR were commonly prescribed at the end of life.

Validation of the Palliative Prognostic Index, Performance Status-Based Palliative Prognostic Index and Chinese Prognostic Scale in a home palliative care setting for patients with advanced cancer in China

BACKGROUND

The predictive value of the prognostic tool for patients with advanced cancer is uncertain in mainland China, especially in the home-based palliative care (HPC) setting. This study aimed to compare the accuracy of the Palliative Prognostic Index (PPI), the Performance Status-Based Palliative Prognostic Index (PS-PPI), and the Chinese Prognosis Scale (ChPS) for patients with advanced cancer in the HPC setting in mainland China.

METHODS

Patients with advanced cancer admitted to the hospice center of Yuebei People's Hospital between January 2014 and December 2018 were retrospectively calculated the scores according to the three prognostic tools. The Kaplan-Meier method was used to compare survival times among different risk groups. Receiver operating characteristic curve analysis was used to assess the predictive value. The accuracy of 21-, 42- and 90-day survival was compared among the three prognostic tools.

RESULTS

A total of 1863 patients were included. Survival time among the risk groups of all prognostic tools was significantly different from each other except for the PPI. The AUROC of the ChPS was significantly higher than that of the PPI and PS-PPI for 7-, 14, 21-, 42-, 90-, 120-, 150- and 180-day survival (P < 0.05). The AUROC of the PPI and PS-PPI were not significantly different from each other (P > 0.05).

CONCLUSIONS

The ChPS is more suitable than the PPI and PS-PPI for advanced cancer patients in the HPC setting. More researches are needed to verify the predictive value of the ChPS, PPI, and PS-PPI in the HPC setting in the future.

Spiritual peace and life meaning may buffer the effect of anxiety on physical well-being in newly diagnosed cancer survivors

OBJECTIVE

To assess the extent to which spiritual well-being moderates the relationship between anxiety and physical well-being in a diverse, community-based cohort of newly diagnosed cancer survivors.

METHODS

Data originated from the Measuring Your Health (MY-Health) study cohort (n = 5506), comprising people assessed within 6-13 months of cancer diagnosis. Life meaning/peace was assessed using the 8-item subscale of the Spiritual Well-Being Scale (FACIT-Sp-12). Anxiety was measured with an 11-item PROMIS Anxiety short form, and physical well-being was assessed using the 7-item FACT-G subscale. Multiple linear regression models were used to assess relationships among variables.

RESULTS

Life meaning and peace was negatively associated with anxiety, b = -0.56 (P < .001) and positively associated with physical well-being, b = 0.43 (P = <.001) after adjusting for race, education, income, and age. A significant interaction between life meaning/peace and anxiety emerged (P < .001) indicating that spiritual well-being moderates the relationship between anxiety and physical well-being. Specifically, for cancer survivors high in anxiety, physical well-being was dependent on levels of life meaning/peace, b = 0.19, P < .001. For those low in anxiety, physical well-being was not associated with levels of life meaning/peace, b = 0.01, P = .541. Differences in cancer clinical factors (cancer stage at diagnosis, cancer type) did not significantly impact results.

CONCLUSIONS

Further research is needed to assess how spiritual well-being may buffer the negative effect of anxiety on physical well-being. A clinical focus on spiritual well-being topics such as peace and life meaning may help cancer survivors of all types as they transition into follow-up care.

Examination of a Short-Term, Prognostic Predictive Method for Terminal Cancer Patients Using the Barthel Index

For the estimation of short-term prognosis in terminal cancer patients, it is important to establish a prognostic index that does not involve blood tests. We compared the prognostic ability of the Barthel Index (BI) with the Glasgow Prognostic Score (GPS). Ninety-seven inpatients with terminal cancer at Onomichi Municipal Hospital who died between 2018 and 2019 were retrospectively analyzed. The sensitivity, specificity, accuracy, and area under the receiver operating characteristic curve (AUROC) were compared between the BI and GPS. For predicting the 15 day prognosis, the BI showed higher specificity, accuracy, and AUROC than the GPS. For predicting the 30 day prognosis, the BI showed higher sensitivity, accuracy, and AUROC than the GPS. The BI can predict the 15 or 30 day prognosis in terminal cancer patients. As the BI does not require blood tests, it may be an option for prognostic prediction in terminal cancer patients.

Prognostic performance of proteomic testing in advanced non-small cell lung cancer: a systematic literature review and meta-analysis

OBJECTIVE

Timely assessment of patient-specific prognosis is critical to oncology care involving a shared decision-making approach, but clinical prognostic factors traditionally used in NSCLC have limitations. We examine a proteomic test to address these limitations.

METHODS

This study examines the prognostic performance of the VeriStrat blood-based proteomic test that measures the inflammatory disease state of patients with advanced NSCLC. A systematic literature review (SLR) was performed, yielding cohorts in which the hazard ratio (HR) was reported for overall survival (OS) of patients with VeriStrat Poor (VSPoor) test results versus VeriStrat Good (VSGood). A study-level meta-analysis of OS HRs was performed in subgroups defined by lines of therapy and treatment regimens.

RESULTS

Twenty-four cohorts met SLR criteria. Meta-analyses in five subgroups (first-line platinum-based chemotherapy, second-line single-agent chemotherapy, first-line EGFR-tyrosine kinase inhibitor (TKI) therapy, and second- and higher-line TKI therapy, and best supportive care) resulted in statistically significant (p ≤ .001) summary effect sizes for OS HRs of 0.42, 0.54, 0.41, 0.52, and 0.50, respectively, indicating increased OS by about two-fold for patients who test VSGood. No significant heterogeneity was seen in any subgroup (p > .05).

CONCLUSIONS

Advanced NSCLC patients classified VSGood have significantly longer OS than those classified VSPoor. The summary effect size for OS HRs around 0.4-0.5 indicates that the expected median survival of those with a VSGood classification is approximately 2-2.5 times as long as those with VSPoor. The robust prognostic performance of the VeriStrat test across various lines of therapy and treatment regimens has clinical implications for treatment shared decision-making and potential for novel treatment strategies.

How palliative care professionals deal with predicting life expectancy at the end of life: predictors and accuracy

PURPOSE

To assess the accuracy of hospice staff in predicting survival of subjects admitted to hospice, exploring the factors considered most helpful by the hospice staff to accurately predict survival.

METHODS

Five physicians and 11 nurses were asked to predict survival at admission of 827 patients. Actual and predicted survival times were divided into ≤ 1 week, 2-3 weeks, 4-8 weeks, and ≥ 2 months and the accuracy of the estimates was calculated. The staff members were each asked to score 17 clinical variables that guided them in predicting survival and we analyzed how these variables impacted the accuracy.

RESULTS

Physicians' and nurses' accuracy of survival of the patients was 46% and 40% respectively. Survival was underestimated in 20% and 12% and overestimated in 34% and 48% of subjects. Both physicians and nurses considered metastases, comorbidities, dyspnea, disability, tumor site, neurological symptoms, and confusion very important in predicting patients' survival with nurses assigning more importance to intestinal symptoms and pain too. All these factors, with the addition of cough and/or bronchial secretions, were associated with physicians' greater accuracy. In the multivariable models, intestinal symptoms and confusion continued to be associated with greater predictive accuracy. No factors appreciably raised nurses' accuracy.

CONCLUSIONS

Some clinical symptoms rated as relevant by the hospice staff could be important for predicting survival. However, only intestinal symptoms and confusion significantly improved the accuracy of physicians' predictions, despite the high prevalence of overestimated survival.

Prognostic Markers of Overall Survival in Cancer Patients Attending a Cachexia Support Service: An Evaluation of Clinically Assessed Physical Function, Malnutrition and Inflammatory Status

BACKGROUND

Cancer cachexia is a muscle-wasting syndrome that results in physical function impairments and decreased survival. While body weight and muscle mass loss predict survival, the prognostic significance of physical function in this population is unclear. Thus, we evaluated the association between physical function, and other routine measures, and overall survival (OS) in cancer patients attending a cachexia support service.

METHODS

Physical function was clinically-assessed using the 30 s sit-to-stand test and handgrip strength. Six-month weight loss, the Patient-Generated Subjective Global Assessment (PG-SGA) total score, C-reactive protein (CRP), albumin, and quality of life were also evaluated.

RESULTS

Records from 203 patients (age: 68.6 ± 11.6 years) were included. Handgrip strength did not predict OS. Sit-to-stand repetitions predicted OS in the single variable, but not the multivariable analysis. Multivariable results suggested higher PG-SGA total scores (HR: 1.04, 95% CI: 1.01-1.07), six-month weight loss (HR: 1.02, 95% CI: 1.004-1.04), and elevated CRP (HR: 1.004, 95% CI: 1.0004-1.01) predicted shorter OS. Higher albumin predicted longer OS (HR: 0.93, 95% CI: 0.90-0.97).

CONCLUSION

Six-month weight loss, the PG-SGA total score, CRP, and albumin independently predicted survival, while physical function did not. Functional impairments remain a hallmark of cancer cachexia and the benefit of their routine assessment warrants further exploration, especially in relation to patient quality of life.

Acute palliative care unit-initiated interventions for advanced cancer patients at the end of life: prediction of impending death based on Glasgow Prognostic Score

PURPOSE

Unexpected death occurred in an unexpectedly high proportion of advanced cancer patients in the acute palliative care unit (APCU) setting and associated with fewer signs of impending death. Recognition of patients at high risk of approaching death, especially immediately after admitting APCU among clinicians, can improve the end-of-life trajectory. Our objective was accurate prognostication within a few days of admission.

METHODS

Patients admitted to an APCU of the NTT Medical Center Tokyo, Tokyo, Japan, between April 2009 and December 2016 were retrospectively examined. The Glasgow Prognostic Score (GPS) was optimized with concomitant neutrophilia, lymphocytopenia, thrombocytopenia, anemia, and monocytosis. Kaplan-Meier survival curves were estimated, and independent predictors for 3-day mortality were identified using univariate and multivariate analyses. The sensitivity, specificity, and likelihood ratios (LRs) associated with imminent death were also assessed.

RESULTS

Nine hundred ninety-one patients were included; 52.9% was male. The median age was 72 years. The median survival was 13 days (IQ range 6 to 26), and 11.7% died within 3 days of admission. Significant difference in survival with a GPS of 2 was observed in GPS optimized with concomitant thrombocytopenia, and it was the only significant predictor associated with 3-day mortality (p = 0.004), which had high specificity (> 95%) and high positive LR (> 5).

CONCLUSION

The prognostic value of the GPS was enhanced by adding thrombocytopenia. The concurrent use of the GPS and platelet count improved the prognostication of limited time of survival and could assist in the personal and clinical decisions for advanced cancer patients.

Patient-reported symptoms and problems at admission to specialized palliative care improved survival prediction in 30,969 cancer patients: A nationwide register-based study

BACKGROUND

Large, nationally representative studies of the association between quality of life and survival time in cancer patients in specialized palliative care are missing.

AIM

The aim of this study was to investigate whether symptoms/problems at admission to specialized palliative care were associated with survival and if the symptoms/problems may improve prediction of death within 1 week and 1 month, respectively.

SETTING/PARTICIPANTS

All cancer patients who had filled in the EORTC QLQ-C15-PAL at admission to specialized palliative care in Denmark in 2010-2017 were included through the Danish Palliative Care Database. Cox regression was used to identify clinical variables (gender, age, type of contact (inpatient vs outpatient), and cancer site) and symptoms/problems significantly associated with survival. To test whether symptoms/problems improved survival predictions, the overall accuracy (area under the receiver operating characteristic curve) for different prediction models was compared. The validity of the prediction models was tested with data on 5,508 patients admitted to palliative care in 2018.

RESULTS

The study included 30,969 patients with an average age of 68.9 years; 50% were women. Gender, age, type of contact, cancer site, and most symptoms/problems were significantly associated with survival time. The predictive value of symptoms/problems was trivial except for physical function, which clearly improved the overall accuracy for 1-week and 1-month predictions of death when added to models including only clinical variables.

CONCLUSION

Most symptoms/problems were significantly associated with survival and mainly physical function improved predictions of death. Interestingly, the predictive value of physical function was the same as all clinical variables combined (in hospice) or even higher (in palliative care teams).

Predicting in-hospital mortality of patients with febrile neutropenia using machine learning models

BACKGROUND

Febrile neutropenia (FN) has been associated with high mortality among adults with cancer. Current systems for early detection of inpatient FN mortality are based on scoring indexes that require intensive physicians' subjective evaluation.

OBJECTIVE

In this study, we leveraged machine learning techniques to build a FN mortality risk evaluation tool focused on FN admissions without physicians' subjective evaluation.

METHODS

We used the National Inpatient Sample and Nationwide Inpatient Sample (NIS) that included mortality data among adult inpatients who were diagnosed with FN during a hospital admission. Machine learning techniques that we compared included linear models (ridge logistic regression and linear support vector machine) and non-linear models (gradient boosting tree and neural network). The primary outcome for this study was death among individuals with a recorded FN admission. Model comparison was evaluated based on areas under the receiver operating characteristic curve (AUROC) and model performance was estimated using 30 % test set created via stratified split.

RESULTS

Our analysis detected 126,013 adult admissions within the NIS data that were diagnosed with FN, among which 5,856 were declared as deceased (4.6 %). Our machine learning results demonstrate linear models and non-linear models achieved areas under the receiver operating characteristic (AUROC) around 92 % in survival prediction.

CONCLUSIONS

We developed machine learning models that do not require physicians' subjective evaluation for FN mortality risk prediction.

Prognosticating for Adult Patients With Advanced Incurable Cancer: a Needed Oncologist Skill

Patients with advanced cancer and their families commonly seek information about prognosis to aid decision-making in medical (e.g. surrounding treatment), psychological (e.g. saying goodbye), and social (e.g. getting affairs in order) domains. Oncologists therefore have a responsibility to identify and address these requests by formulating and sensitively communicating information about prognosis. Current evidence suggests that clinician predictions are correlated with actual survival but tend to be overestimations. In an attempt to cultivate prognostic skills, it is recommended that clinicians practice formulating and recording subjective estimates of prognosis in advanced cancer patient’s medical notes. When possible, a multi-professional prognostic estimate should be sought as these may be more accurate than individual predictions alone. Clinicians may consider auditing the accuracy of their predictions periodically and using feedback from this process to improve their prognostic skills.

Clinicians may also consider using validated prognostic tools to complement their clinical judgements. However, there is currently only limited evidence about the comparative accuracy of different prognostic tools or the extent to which these measures are superior to clinical judgement. Oncologists and palliative care physicians should ensure that they receive adequate training in advanced communication skills, which builds upon their pre-existing skills, to sensitively deliver information on prognosis. In particular, clinicians should acknowledge their own prognostic uncertainty and should emphasise the supportive care that can continue to be provided after stopping cancer-directed therapies.

Chemotherapy-induced neuroinflammation is associated with disrupted colonic and bacterial homeostasis in female mice

Chemotherapy treatment negatively affects the nervous and immune systems and alters gastrointestinal function and microbial composition. Outside of the cancer field, alterations in commensal bacteria and immune function have been implicated in behavioral deficits; however, the extent to which intestinal changes are related to chemotherapy-associated behavioral comorbidities is not yet known. Thus, this study identified concurrent changes in behavior, central and peripheral immune activation, colon histology, and bacterial community structure in mice treated with paclitaxel chemotherapy. In paclitaxel-treated mice, increased fatigue and decreased cognitive performance occurred in parallel with reduced microglia immunoreactivity, increased circulating chemokine expression (CXCL1), as well as transient increases in pro-inflammatory cytokine/chemokine (Il-1β, Tnfα, Il-6, and Cxcl1) gene expression in the brain. Furthermore, mice treated with paclitaxel had altered colonic bacterial community composition and increased crypt depth. Relative abundances of multiple bacterial taxa were associated with paclitaxel-induced increases in colon mass, spleen mass, and microglia activation. Although microbial community composition was not directly related to available brain or behavioral measures, structural differences in colonic tissue were strongly related to microglia activation in the dentate gyrus and the prefrontal cortex. These data indicate that the chemotherapeutic paclitaxel concurrently affects the gut microbiome, colonic tissue integrity, microglia activation, and fatigue in female mice, thus identifying a novel relationship between colonic tissue integrity and behavioral responses that is not often assessed in studies of the brain-gut-microbiota axis.

Translational and basic science opportunities in palliative care and radiation oncology

Radiation therapy is commonly used in the metastatic setting to palliate pain, neurological deficits, bleeding and other complications of metastatic disease, allowing patients to live longer and have better quality of life. Despite the effective use of radiation and other palliative treatment modalities, many patients continue to experience poorly controlled pain and other serious sequelae of their disease, underscoring the need for additional research in this area. In this review we highlight recent developments impacting the fields of palliative care and radiation oncology and describe opportunities for research and innovation including studies of tumor microenvironment, identification of effective biomarkers of tumor response and combinatorial treatments with new systemic agents. It is our hope that progress in these fields will improve the lives of patients living with advanced malignancies.

Improving palliative and end-of-life care with machine learning and routine data: a rapid review

Introduction: Improving end-of-life (EOL) care is a priority worldwide as this population experiences poor outcomes and accounts disproportionately for costs. In clinical practice, physician judgement is the core method of identifying EOL care needs but has important limitations. Machine learning (ML) is a subset of artificial intelligence advancing capacity to identify patterns and make predictions using large datasets.  ML approaches have the potential to improve clinical decision-making and policy design, but there has been no systematic assembly of current evidence. Methods: We conducted a rapid review, searching systematically seven databases from inception to December 31st, 2018: EMBASE, MEDLINE, Cochrane Library, PsycINFO, WOS, SCOPUS and ECONLIT.  We included peer-reviewed studies that used ML approaches on routine data to improve palliative and EOL care for adults.  Our specified outcomes were survival, quality of life (QoL), place of death, costs, and receipt of high-intensity treatment near end of life.  We did not search grey literature and excluded material that was not a peer-reviewed article. Results: The database search identified 426 citations. We discarded 162 duplicates and screened 264 unique title/abstracts, of which 22 were forwarded for full text review.  Three papers were included, 18 papers were excluded and one full text was sought but unobtainable.  One paper predicted six-month mortality, one paper predicted 12-month mortality and one paper cross-referenced predicted 12-month mortality with healthcare spending.  ML-informed models outperformed logistic regression in predicting mortality but poor prognosis is a weak driver of costs.  Models using only routine administrative data had limited benefit from ML methods. Conclusion: While ML can in principle help to identify those at risk of adverse outcomes and inappropriate treatment near EOL, applications to policy and practice are formative.  Future research must not only expand scope to other outcomes and longer timeframes, but also engage with individual preferences and ethical challenges.

Improving palliative and end-of-life care with machine learning and routine data: a rapid review

Introduction: Improving end-of-life (EOL) care is a priority worldwide as this population experiences poor outcomes and accounts disproportionately for costs. In clinical practice, physician judgement is the core method of identifying EOL care needs but has important limitations. Machine learning (ML) is a subset of artificial intelligence advancing capacity to identify patterns and make predictions using large datasets.  ML approaches have the potential to improve clinical decision-making and policy design, but there has been no systematic assembly of current evidence. Methods: We conducted a rapid review, searching systematically seven databases from inception to December 31st, 2018: EMBASE, MEDLINE, Cochrane Library, PsycINFO, WOS, SCOPUS and ECONLIT.  We included peer-reviewed studies that used ML approaches on routine data to improve palliative and EOL care for adults.  Our specified outcomes were survival, quality of life (QoL), place of death, costs, and receipt of high-intensity treatment near end of life.  We did not search grey literature and excluded material that was not a peer-reviewed article. Results: The database search identified 426 citations. We discarded 162 duplicates and screened 264 unique title/abstracts, of which 22 were forwarded for full text review.  Three papers were included, 18 papers were excluded and one full text was sought but unobtainable.  One paper predicted six-month mortality, one paper predicted 12-month mortality and one paper cross-referenced predicted 12-month mortality with healthcare spending.  ML-informed models outperformed logistic regression in predicting mortality but poor prognosis is a weak driver of costs.  Models using only routine administrative data had limited benefit from ML methods. Conclusion: While ML can in principle help to identify those at risk of adverse outcomes and inappropriate treatment near EOL, applications to policy and practice are formative.  Future research must not only expand scope to other outcomes and longer timeframes, but also engage with individual preferences and ethical challenges.

Prediction of early mortality in patients with cancer-associated thrombosis in the RIETE Database

BACKGROUND

Proper risk stratification of patients for early mortality after cancer-associated thrombosis may lead to personalized anticoagulation protocols. Therefore, we aimed to derive and validate a scoring system to predict early mortality in this population. To this end, we selected patients with active cancer and thrombosis from the Computerized Registry of Patients with Venous Thromboembolism database.

METHODS

The main outcome was all cause mortality within the month following a thrombotic event. We used a simple random selection to split are data in a derivation and a validation cohort. In the derivation cohort, we used recursive partitioning and binary logistic regression to identify groups at risk and to determine the likelihood of the primary outcome. The risk score was developed based on odds ratios from the final multivariate model, and then tested in the validation cohort.

RESULTS

In 10,025 eligible patients, we identified 6 predictors of 30-day mortality: leukocytosis ≥11.5x109/L; platelet count ≤160x109/L, metastasis, recent immobility, initial presentation as pulmonary embolism and Body Mass Index <18.5. The model divided the population into 3 risk categories: low (score 0-3), moderate (score 4-6), and high (score ≥7). The AUC for the overall score was 0.74, and using a cutoff ≥7 points, the model had a negative predictive value of 94.4%, a positive predictive value of 23.1%, a sensitivity of 73.3%, and a specificity of 64.6% in the validation cohort.

CONCLUSIONS

Our validated risk model may assist physicians in the selection of patients for outpatient management, and perhaps anticoagulant, considering expanding anticoagulation options.

Prognostic accuracy of patients, caregivers, and oncologists in advanced cancer

BACKGROUND

In caring for patients with advanced cancer, accurate estimation of survival is important for clinical decision making. The purpose of this study was to assess the accuracy of 2-year survival probabilities estimated by oncologists, patients, and caregivers and to identify demographic and clinical factors associated with prognostic accuracy.

METHODS

This was a secondary observational analysis of data obtained from a cluster randomized controlled trial. Participants included 38 oncologists, 263 patients with advanced nonhematologic cancer, and 193 of their caregivers from clinics in Sacramento and Western New York. Discrimination within each group (oncologists, patients, caregivers) was evaluated using the C statistic, whereas calibration was assessed by comparing observed to predicted 2-year mortality using the chi-square statistic.

RESULTS

The median survival from study entry was 18 months, and 41.8% of patients survived for 2 years. C statistics for oncologists, patients, and caregivers were 0.81 (95% CI, 0.76-0.86), 0.62 (95% CI, 0.55-0.68), and 0.72 (95% CI, 0.65-0.78), respectively; oncologists' predictions were better than the predictions of both patients (P = .001) and caregivers (P = .03). Oncologists also had superior calibration: their predictions of 2-year survival were similar to actual survival (P = .17), whereas patients' (P = .0001) and caregivers' (P = .003) predictions diverged significantly from actual survival. Although most oncologists' predictions were classified as realistic (62.0%), approximately one-half of patients' and caregivers' predictions (50.0% and 46.0%, respectively) were unduly optimistic. Among patients, nonwhite race and higher levels of social well-being predicted undue optimism (P < .05).

CONCLUSIONS

Compared with oncologists, patients and caregivers displayed inferior prognostic discrimination, and their predictions were poorly calibrated, primarily because of overoptimism.

Accuracy and Prognostic Significance of Oncologists' Estimates and Scenarios for Survival Time in Advanced Gastric Cancer

BACKGROUND

Worst-case, typical, and best-case scenarios for survival, based on simple multiples of an individual's expected survival time (EST), estimated by their oncologist, are a useful way of formulating and explaining prognosis. We aimed to determine the accuracy and prognostic significance of oncologists' estimates of EST, and the accuracy of the resulting scenarios for survival time, in advanced gastric cancer.

MATERIALS AND METHODS

Sixty-six oncologists estimated the EST at baseline for each of the 152 participants they enrolled in the INTEGRATE trial. We hypothesized that oncologists' estimates of EST would be unbiased (∼50% would be longer or shorter than the observed survival time [OST]); imprecise (<33% within 0.67-1.33 times the OST); independently predictive of overall survival (OS); and accurate at deriving scenarios for survival time with approximately 10% of patients dying within a quarter of their EST (worst-case scenario), 50% living within half to double their EST (typical scenario), and 10% living three or more times their EST (best-case scenario).

RESULTS

Oncologists' estimates of EST were unbiased (45% were shorter than the OST, 55% were longer); imprecise (29% were within 0.67-1.33 times observed); moderately discriminative (Harrell's C-statistic 0.62, p = .001); and an independently significant predictor of OS (hazard ratio, 0.89; 95% confidence interval, 0.83-0.95; p = .001) in a Cox model including performance status, number of metastatic sites, neutrophil-to-lymphocyte ratio ≥3, treatment group, age, and health-related quality of life (EORTC-QLQC30 physical function score). Scenarios for survival time derived from oncologists' estimates were remarkably accurate: 9% of patients died within a quarter of their EST, 57% lived within half to double their EST, and 12% lived three times their EST or longer.

CONCLUSION

Oncologists' estimates of EST were unbiased, imprecise, moderately discriminative, and independently significant predictors of OS. Simple multiples of the EST accurately estimated worst-case, typical, and best-case scenarios for survival time in advanced gastric cancer.

IMPLICATIONS FOR PRACTICE

Results of this study demonstrate that oncologists' estimates of expected survival time for their patients with advanced gastric cancer were unbiased, imprecise, moderately discriminative, and independently significant predictors of overall survival. Simple multiples of the expected survival time accurately estimated worst-case, typical, and best-case scenarios for survival time in advanced gastric cancer.

Palliative radiotherapy near the end of life

Background

A significant proportion of patients with advanced cancer undergo palliative radiotherapy (RT) within their last 30 days of life. This study characterizes palliative RT at our institution and aims to identify patients who may experience limited benefit from RT due to imminent mortality.

Methods

Five hundred and-eighteen patients treated with external beam RT to a site of metastatic disease between 2012 and 2016 were included. Mann-Whitney U and chi-squared tests were used to identify factors associated with RT within 30 days of death (D₃₀RT).

Results

Median age at RT was 63 years (IQR 54–71). Median time from RT to death was 74 days (IQR 33–174). One hundred and twenty-five patients (24%) died within 30 days of RT. D₃₀RT was associated with older age at RT (64 vs. 62 years, p = 0.04), shorter interval since diagnosis (14 vs. 31 months, p <  0.001), liver metastasis (p = 0.02), lower KPS (50 vs. 70, p <  0.001), lower BMI (22 vs. 24, p = 0.001), and inpatient status at consult (56% vs. 26%, p < 0.001). Patients who died within 30 days of RT were less likely to have hospice involved in their care (44% vs. 71%, p = 0.001). D₃₀RT was associated with higher Chow and TEACHH scores at consult (p < 0.001 for both).

Conclusions

Twenty-four percent of patients received palliative RT within 30 days of death. Additional tools are necessary to help physicians identify patients who would benefit from short treatment courses or alternative interventions to maximize quality at the end of life.