Modeling recurrence in colorectal cancer

Colorectal cancer recurrence and its impact on survival after curative surgery: An analysis based on multistate models

AIM

To investigate the usefulness of multistate models (MSM) for determining colorectal cancer (CRC) recurrence rate, to analyse the effect of different factors on tumour recurrence and death, and to assess the impact of recurrence for CRC prognosis.

METHODS

Observational follow-up study of incident CRC cases disease-free after curative resection in 2006-2013 (n = 994). Recurrence and mortality were analyzed with MSM, as well as covariate effects on transition probabilities.

RESULTS

Cumulative incidence of recurrence at 60 months was 13.7%. Five years after surgery, 70.3% of patients were alive and recurrence-free, and 8.4% were alive after recurrence. Recurrence has a negative impact on prognosis, with 5-year CRC-related mortality increasing from 3.8% for those who are recurrence-free 1-year after surgery to 33.6% for those with a recurrence. Advanced stage increases recurrence risk (HR = 1.53) and CRC-related mortality after recurrence (HR = 2.35). CRC-related death was associated with age in recurrence-free patients, and with comorbidity after recurrence. As expected, age≥75 years was a risk factor for non-CRC-related death with (HR = 7.76) or without recurrence (HR = 4.26), while its effect on recurrence risk was not demonstrated.

CONCLUSIONS

MSM allows detailed analysis of recurrence and mortality in CRC. Recurrence has a negative impact on prognosis. Advanced stage was a determining factor for recurrence and CRC-death after recurrence.

Exploring the impact of stage and tumor site on colorectal cancer survival: Bayesian survival modeling

Colorectal cancer is a prevalent malignancy with global significance. This retrospective study aimed to investigate the influence of stage and tumor site on survival outcomes in 284 colorectal cancer patients diagnosed between 2001 and 2017. Patients were categorized into four groups based on tumor site (colon and rectum) and disease stage (early stage and advanced stage). Demographic characteristics, treatment modalities, and survival outcomes were recorded. Bayesian survival modeling was performed using semi-competing risks illness-death models with an accelerated failure time (AFT) approach, utilizing R 4.1 software. Results demonstrated significantly higher time ratios for disease recurrence (TR = 1.712, 95% CI 1.489-2.197), mortality without recurrence (TR = 1.933, 1.480-2.510), and mortality after recurrence (TR = 1.847, 1.147-2.178) in early-stage colon cancer compared to early-stage rectal cancer. Furthermore, patients with advanced-stage rectal cancer exhibited shorter survival times for disease recurrence than patients with early-stage colon cancer. The interaction effect between the disease site and cancer stage was not significant. These findings, derived from the optimal Bayesian log-normal model for terminal and non-terminal events, highlight the importance of early detection and effective management strategies for colon cancer. Early-stage colon cancer demonstrated improved survival rates for disease recurrence, mortality without recurrence, and mortality after recurrence compared to other stages. Early intervention and comprehensive care are crucial to enhance prognosis and minimize adverse events in colon cancer patients.

A bayesian approach to model the underlying predictors of early recurrence and postoperative death in patients with colorectal Cancer

Objective

This study aimed at utilizing a Bayesian approach semi-competing risks technique to model the underlying predictors of early recurrence and postoperative Death in patients with colorectal cancer (CRC).

Methods

In this prospective cohort study, 284 patients with colorectal cancer, who underwent surgery, referred to Imam Khomeini clinic in Hamadan from 2001 to 2017. The primary outcomes were the probability of recurrence, the probability of Mortality without recurrence, and the probability of Mortality after recurrence. The patients ‘recurrence status was determined from patients’ records. The Bayesian survival modeling was carried out by semi-competing risks illness-death models, with accelerated failure time (AFT) approach, in R 4.1 software. The best model was chosen according to the lowest deviance information criterion (DIC) and highest logarithm of the pseudo marginal likelihood (LPML).

Results

The log-normal model (DIC = 1633, LPML = -811), was the optimal model. The results showed that gender(Time Ratio = 0.764: 95% Confidence Interval = 0.456–0.855), age at diagnosis (0.764: 0.538–0.935 ), T₃ stage (0601: 0.530–0.713), N₂ stage (0.714: 0.577–0.935 ), tumor size (0.709: 0.610–0.929), grade of differentiation at poor (0.856: 0.733–0.988), and moderate (0.648: 0.503–0.955) levels, and the number of chemotherapies (1.583: 1.367–1.863) were significantly related to recurrence. Also, age at diagnosis (0.396: 0.313–0.532), metastasis to other sites (0.566: 0.490–0.835), T₃ stage (0.363: 0.592 − 0.301), T₄ stage (0.434: 0.347–0.545), grade of differentiation at moderate level (0.527: 0.387–0.674), tumor size (0.595: 0.500–0.679), and the number of chemotherapies (1.541: 1.332–2.243) were the significantly predicted the death. Also, age at diagnosis (0.659: 0.559–0.803), and the number of chemotherapies (2.029: 1.792–2.191) were significantly related to mortality after recurrence.

Conclusion

According to specific results obtained from the optimal Bayesian log-normal model for terminal and non-terminal events, appropriate screening strategies and the earlier detection of CRC leads to substantial improvements in the survival of patients.

Development and Validation of a Decision Analytical Model for Posttreatment Surveillance for Patients With Oropharyngeal Carcinoma

IMPORTANCE

Clinical practice regarding posttreatment radiologic surveillance for patients with oropharyngeal carcinoma (OPC) is neither adapted to individual patient risk nor fully evidence based.

OBJECTIVES

To construct a microsimulation model for posttreatment OPC progression and use it to optimize surveillance strategies while accounting for both tumor stage and human papillomavirus (HPV) status.

DESIGN, SETTING, AND PARTICIPANTS

In this decision analytical modeling study, a Markov model of 3-year posttreatment patient trajectories was created. The training data source was the American College of Surgeon's National Cancer Database from 2010 to 2015. The external validation data set was the 2016 International Collaboration on Oropharyngeal Cancer Network for Staging (ICON-S) study. Training data comprised 2159 patients with OPC treated with primary radiotherapy who had known HPV status and disease staging information. Patients with American Joint Committee on Cancer, 7th edition stage III to IVB disease and those with clinical metastases during the time of primary treatment were included. Data were analyzed from August 1 to October 31, 2020.

MAIN OUTCOMES AND MEASURES

Main outcomes included disease stage and HPV status, specific disease transition probabilities, and latency of surveillance regimens, defined as time between recurrence incidence and disease discovery.

RESULTS

Training data consisted of 2159 total patients (1708 men [79.1%]; median age, 59.6 years [range, 40-90 years]; 401 with stage III disease, 1415 with stage IVA disease, and 343 with stage IVB disease). Cohorts predominantly had HPV-negative disease (1606 [74.4%]). With model-optimized regimens, recurrent disease was discovered a mean of 0.6 months (95% CI, 0.5-0.8 months) earlier than with a standard surveillance regimen based on current clinical guidelines. Recurrent disease was discovered using the optimized regimens without significant reduction in sensitivity. Compared with strategies based on reimbursement guidelines, the model-optimized regimens found disease a mean of 1.8 months (95% CI, 1.3-2.3 months) earlier.

CONCLUSIONS AND RELEVANCE

Optimized, risk-stratified surveillance regimens consistently outperformed nonoptimized strategies. These gains were obtained without requiring any additional imaging studies. This approach to risk-stratified surveillance optimization is generalizable to a broad range of tumor types and risk factors.

Joint modelling of colorectal cancer recurrence and death after resection using multi-state model with cured fraction

Curing of colorectal cancer (CRC) occurs at the time of resection but it is not immediately observable. If the cancer is not completely eliminated, the patient will not be cured of cancer and will experience recurrence as the tumor has regrown to a detectable size. The main propose of the present study was to assess the effects of different covariates on the probability of being cured as well as the time-to-recurrence, and time-to-death in CRC patients by using multi-state cure model. The information of 283 patients with CRC, who underwent resection, from 2000 to 2015 in Imam Khomeini Hospital of Hamadan, Iran, were analyzed. The results of multi-state cure model reveal that females and who experience metastasis were more likely to be apparently cured. It has been shown that sex has a significant effect on the time-to-recurrence given patient was in the not cured group. The survival time of patients of the not cured group was affected by the stage of disease. However, the survival of the apparently cured patients were affected by age at diagnosis and metastasis status. The multi-state cure model provided a flexible framework to study the effects of prognostic factors simultaneously on the transition between different states and the probability of being apparently cured of CRC.

Metastatic Potential and Survival of Duodenal and Pancreatic Tumors in Multiple Endocrine Neoplasia Type 1: A GTE and AFCE Cohort Study (Groupe d'étude des Tumeurs Endocrines and Association Francophone de Chirurgie Endocrinienne)

OBJECTIVE

To assess the distant metastatic potential of duodeno-pancreatic neuroendocrine tumors (DP-NETs) in patients with MEN1, according to functional status and size.

SUMMARY BACKGROUND DATA

DP-NETs, with their numerous lesions and endocrine secretion-related symptoms, continue to be a medical challenge; unfortunately they can become aggressive tumors associated with distant metastasis, shortening survival. The survival of patients with large nonfunctional DP-NETs is known to be poor, but the overall contribution of DP-NETs to metastatic spread is poorly known.

METHODS

The study population included patients with DP-NETs diagnosed after 1990 and followed in the MEN1 cohort of the Groupe d'étude des Tumeurs Endocrines (GTE). A multistate Markov piecewise constant intensities model was applied to separate the effects of prognostic factors on 1) metastasis, and 2) metastasis-free death or 3) death after appearance of metastases.

RESULTS

Among the 603 patients included, 39 had metastasis at diagnosis of DP-NET, 50 developed metastases during follow-up, and 69 died. The Markov model showed that Zollinger-Ellison-related tumors (regardless of tumor size and thymic tumor pejorative impact), large tumors over 2 cm, and age over 40 years were independently associated with an increased risk of metastases. Men, patients over 40 years old and patients with tumors larger than 2 cm, also had an increased risk of death once metastasis appeared.

CONCLUSIONS

DP-NETs of 2 cm in size or more, regardless of the associated secretion, should be removed to prevent metastasis and increase survival. Surgery for gastrinoma remains debatable.

A hidden Markov model for population-level cervical cancer screening data

The Cancer Registry of Norway has been administrating a national cervical cancer screening program since 1992 by coordinating triennial cytology exam screenings for the female population between 25 and 69 years of age. Up to 80% of cancers are prevented through mass screening, but this comes at the expense of considerable screening activity and leads to overtreatment of clinically asymptomatic precancers. In this article, we present a continuous-time, time-inhomogeneous hidden Markov model which was developed to understand the screening process and cervical cancer carcinogenesis in detail. By leveraging 1.7 million individual's multivariate time-series of medical exams performed over a 25-year period, we simultaneously estimate all model parameters. We show that an age-dependent model reflects the Norwegian screening program by comparing empirical survival curves from observed registry data and data simulated from the proposed model. The model can be generalized to include more detailed individual-level covariates as well as new types of screening exams. By utilizing individual screening histories and covariate data, the proposed model shows potential for improving strategies for cancer screening programs by personalizing recommended screening intervals.

Impact of delaying initiation of anaplastic lymphoma kinase inhibitor treatment on survival in patients with advanced non-small-cell lung cancer

INTRODUCTION

Several obstacles may delay receipt of targeted treatment in patients with anaplastic lymphoma kinase positive (ALK+) non-small-cell lung cancer (NSCLC). This study examined the factors associated with delayed initiation of ALK inhibitor (ALKi) treatment and its impact on overall survival (OS) as well as the impact of initiating chemotherapy before biomarker test results.

MATERIALS AND METHODS

Advanced NSCLC (aNSCLC) patients selected from the deidentified Flatiron Health electronic health record-derived database were stratified into early- and delayed-use cohorts based on initiation of ALKi treatment relative to time since receiving ALK+ biomarker test results; cohorts were further stratified by timing of chemotherapy initiation relative to availability of ALK+ test results. Prescription-time matching (PTM) was used to examine the effect of delayed ALKi treatment and chemotherapy on survival; Cox proportional hazards models adjusting for baseline characteristics before and after PTM were used to examine factors associated with delayed ALKi treatment and the effects of delayed ALKi treatment and chemotherapy on OS, respectively.

RESULTS

Comparison of OS between early- and delayed-use cohorts (N = 442 ALK + aNSCLC patients) demonstrated that a >3-week delay in the initiation of ALKi treatment was associated with a >2-fold higher risk of death (adjusted hazard ratio [HR] [95 % CI] 2.05 [1.13, 3.71]. The number of office visits, age factors, and use of chemotherapy were associated with an increased risk of being untreated >3 weeks after ALK+ test results. There were no significant differences in survival outcomes regardless of whether patients received chemotherapy before the ALK+ test result or ALKi treatment (adjusted HR [95 % CI] 1.02 [0.64, 1.63]). Completing the chemotherapy regimen after receiving ALK+ test results did not appear to improve survival (adjusted HR [95 % CI] 0.84 [0.38, 1.9]).

CONCLUSION

Initiating ALKi treatment for aNSCLC patients in a timely manner may have a positive impact on survival outcomes.

Multistate Markov Model to Predict the Prognosis of High-Risk Human Papillomavirus-Related Cervical Lesions

Cervical intraepithelial neoplasia (CIN) has a natural history of bidirectional transition between different states. Therefore, conventional statistical models assuming a unidirectional disease progression may oversimplify CIN fate. We applied a continuous-time multistate Markov model to predict this CIN fate by addressing the probability of transitions between multiple states according to the genotypes of high-risk human papillomavirus (HPV). This retrospective cohort comprised 6022 observations in 737 patients (195 normal, 259 CIN1, and 283 CIN2 patients at the time of entry in the cohort). Patients were followed up or treated at the University of Tokyo Hospital between 2008 and 2015. Our model captured the prevalence trend satisfactory, particularly for up to two years. The estimated probabilities for 2-year transition to CIN3 or more were the highest in HPV 16-positive patients (13%, 30%, and 42% from normal, CIN1, and CIN2, respectively) compared with those in the other genotype-positive patients (3.1%-9.6%, 7.6%-16%, and 21%-32% from normal, CIN1, and CIN2, respectively). Approximately 40% of HPV 52- or 58-related CINs remained at CIN1 and CIN2. The Markov model highlights the differences in transition and progression patterns between high-risk HPV-related CINs. HPV genotype-based management may be desirable for patients with cervical lesions.

Anti-tumour activity of everolimus and sunitinib in neuroendocrine neoplasms

Comparisons between everolimus and sunitinib regarding their efficacy and safety in neuroendocrine neoplasms (NENs) are scarce. We retrospectively analysed the clinicopathological characteristics and outcomes in 92 patients with well-differentiated (WD) NEN of different origin (57 pancreatic NENs (PanNENs)), treated with molecular targeted therapy (MTT) with everolimus or sunitinib, first- (73:19) or second-line (sequential; 12:22) for progressive disease. Disease control rates (DCR: partial response or stable disease) at first-line were higher in all patients treated with everolimus than sunitinib (64/73 vs 12/19, P = 0.012). In PanNENs, DCR at first-line everolimus was 36/42 versus 9/15 with sunitinib (P = 0.062). Progression-free survival (PFS) at first-line everolimus was longer than sunitinib (31 months (95% CI: 23.1-38.9) vs 9 months (95% CI: 0-18.5); log-rank P < 0.0001) in the whole cohort and the subset of PanNENs (log-rank P < 0.0001). Median PFS at second-line MTT was 12 months with everolimus (95% CI: 4.1-19.9) vs 13 months with sunitinib (95% CI: 9.3-16.7; log-rank P = 0.951). Treatment with sunitinib (HR: 3.47; 95% CI: 1.5-8.3; P value: 0.005), KI67 >20% (HR: 6.38; 95% CI: 1.3-31.3; P = 0.022) and prior chemotherapy (HR: 2.71; 95% CI: 1.2-6.3; P = 0.021) were negative predictors for PFS at first line in multivariable and also confirmed at multi-state modelling analyses. Side effect (SE) analysis indicated events of serious toxicities (Grades 3 and 4: n = 13/85 for everolimus and n = 4/41 for sunitinib). Discontinuation rate due to SEs was 20/85 for everolimus versus 4/41 for sunitinib (P = 0.065). No additive toxicity of second-line MTT was confirmed. Based on these findings, and until reliable predictors of response become available, everolimus may be preferable to sunitinib when initiating MTT in progressive NENs.

Estimation in the progressive illness-death model: A nonexhaustive review

Multistate models can be successfully used for describing complex event history data, for example, describing stages in the disease progression of a patient. The so-called "illness-death" model plays a central role in the theory and practice of these models. Many time-to-event datasets from medical studies with multiple end points can be reduced to this generic structure. In these models one important goal is the modeling of transition rates but biomedical researchers are also interested in reporting interpretable results in a simple and summarized manner. These include estimates of predictive probabilities, such as the transition probabilities, occupation probabilities, cumulative incidence functions, and the sojourn time distributions. We will give a review of some of the available methods for estimating such quantities in the progressive illness-death model conditionally (or not) on covariate measures. For some of these quantities estimators based on subsampling are employed. Subsampling, also referred to as landmarking, leads to small sample sizes and usually to heavily censored data leading to estimators with higher variability. To overcome this issue estimators based on a preliminary estimation (presmoothing) of the probability of censoring may be used. Among these, the presmoothed estimators for the cumulative incidences are new. We also introduce feasible estimation methods for the cumulative incidence function conditionally on covariate measures. The proposed methods are illustrated using real data. A comparative simulation study of several estimation approaches is performed and existing software in the form of R packages is discussed.

Multi-state relative survival modelling of colorectal cancer progression and mortality

Accurate identification of factors associated with progression of colorectal cancer remains a challenge. In particular, it is unclear which statistical methods are most suitable to separate the effects of putative prognostic factors on cancer progression vs cancer-specific and other cause mortality. To address these challenges, we analyzed 10 year follow-up data for patients who underwent curative surgery for colorectal cancer in 1985-2000. Separate analyses were performed in two French cancer registries. Results of three multivariable models were compared: Cox model with recurrence as a time-dependent variable, and two multi-state models, which separated prognostic factor effects on recurrence vs death, with or without recurrence. Conventional multi-state model analyzed all-cause mortality while new relative survival multi-state model focused on cancer-specific mortality. Among the 2517 and 2677 patients in the two registries, about 50% died without a recurrence, and 28% had a recurrence, of whom almost 90% died. In both multi-state models men had significantly increased risk of cancer recurrence in both registries (HR=0.79; 95% CI: 0.68-0.92 and HR=0.83; 95% CI: 0.71-0.96). However, the two multi-state models identified different prognostic factors for mortality without recurrence. In contrast to the conventional model, in the relative survival analyses gender had no independent association with cancer-specific mortality whereas patients diagnosed with stage III cancer had significantly higher risks in both registries (HR=1.67; 95% CI: 1.27-2.22 and HR=2.38; 95% CI: 1.29-3.27). In conclusion, relative survival multi-state model revealed that different factors may be associated with cancer recurrence vs cancer-specific mortality either after or without a recurrence.

Modelling recurrent events: a tutorial for analysis in epidemiology

In many biomedical studies, the event of interest can occur more than once in a participant. These events are termed recurrent events. However, the majority of analyses focus only on time to the first event, ignoring the subsequent events. Several statistical models have been proposed for analysing multiple events. In this paper we explore and illustrate several modelling techniques for analysis of recurrent time-to-event data, including conditional models for multivariate survival data (AG, PWP-TT and PWP-GT), marginal means/rates models, frailty and multi-state models. We also provide a tutorial for analysing such type of data, with three widely used statistical software programmes. Different approaches and software are illustrated using data from a bladder cancer project and from a study on lower respiratory tract infection in children in Brazil. Finally, we make recommendations for modelling strategy selection for analysis of recurrent event data.

Health-related quality of life and risk of colorectal cancer recurrence and All-cause death among advanced stages of colorectal cancer 1-year after diagnosis

Background

The study aimed to examine the association between health-related quality of life (HRQOL) assessed with overall survival (OS) and recurrence after diagnosis of colorectal cancer (CRC).

Methods

Overall 160 patients with advanced stage CRC were recruited in an observational study and completed the generic and condition-specific HRQOL questionnaires at the colorectal specialist outpatient clinic in Hong Kong, between 10/2009 and 07/2010. Socio-demographic and clinical characteristics including duration since diagnosis, primary tumor location and treatment modality, were collected to serve as predictor variables in regression models. All-cause death or CRC recurrence was the event of interest. Association between HRQOL with OS was assessed using Cox regression. Association between HRQOL and CRC recurrence was further modeled by competing-risks regression adjusted for the competing-risks of death from any cause.

Results

After a median follow-up of 23 months, there were 22 (16.1%) incidents of CRC recurrence and 15 (9.4%) deaths. Decreased physical functioning (hazard ratios, HR = 0.917, 95% CI:0.889-0.981) and general health of domains in SF-12 (HR = 0.846, 95% CI:0.746-0.958) or SF-6D scores (HR = 0.010, 95% CI:0.000-0.573) were associated with an increased risk of death, with adjustment of patients’ characteristics. Increased vitality (HR = 1.151, 95% CI:1.027-1.289) and mental health (HR = 1.128, 95% CI:1.005-1.265) were associated with an increased likelihood of death. In models adjusted for competing-risk of death, those with worse HRQOL was not associated with increased risk of CRC recurrence.

Conclusions

Although self-reported HRQOL was not a significant prognostic factor for CRC recurrence, the HRQOL provided independent prognostic value about mortality in patients with advanced stage of CRC.

Timing of complications and length of stay after rectal cancer surgery

BACKGROUND

Enhanced recovery pathways have been shown to improve short-term outcomes after colorectal surgery. Occurrence of complications can lead to prolonged length of stay (LOS). The goal of this study was to examine whether shorter time to occurrence of complications was associated with a shorter hospital LOS in rectal cancer patients undergoing minimally invasive surgery, taking into account the perioperative pathway.

STUDY DESIGN

This retrospective study included consecutive patients undergoing rectal cancer resection from 2005 to 2011 at a single institution. Enhanced recovery pathway was introduced in 2009. Complications and date of occurrence were reviewed. The impact of perioperative care modalities and comorbidities was evaluated using competing risk models with occurrence of complications and LOS as time-dependent outcomes measured as time from surgery.

RESULTS

A total of 346 patients were included in the analysis with 78 patients treated with enhanced recovery pathway, and 268 with established care. The overall complication rate was 22.3% (77 patients with ileus, wound infection, leak, abscess, small bowel obstruction, reoperation for bleeding, and renal failure). Median time to occurrence of a complication was 3 days post operation. The time to complication diagnosis was associated with shorter time to discharge after the advent of the complication (hazard ratio = 0.84; 95% CI, 0.73-0.96; p = 0.01). Enhanced recovery pathway was associated with a shorter LOS for patients without complications compared with the established pathway (hazard ratio = 2.81; 95% CI, 2.09-3.78; p < 0.001) after adjusting for comorbidities in a competing risk model.

CONCLUSIONS

Early diagnosis of postoperative complications is associated with a shorter LOS after rectal cancer surgery. Enhanced recovery pathway can facilitate a faster recovery in the presence of comorbidities.

Polyisoprenylated methylated protein methyl esterase is both sensitive to curcumin and overexpressed in colorectal cancer: implications for chemoprevention and treatment

Inhibition of PMPMEase, a key enzyme in the polyisoprenylation pathway, induces cancer cell death. In this study, purified PMPMEase was inhibited by the chemopreventive agent, curcumin, with a K(i) of 0.3 μM (IC50 = 12.4 μM). Preincubation of PMPMEase with 1 mM curcumin followed by gel-filtration chromatography resulted in recovery of the enzyme activity, indicative of reversible inhibition. Kinetics analysis with N-para-nitrobenzoyl-S-trans,trans-farnesylcysteine methyl ester substrate yielded K M values of 23.6 ± 2.7 and 85.3 ± 15.3 μM in the absence or presence of 20 μM curcumin, respectively. Treatment of colorectal cancer (Caco2) cells with curcumin resulted in concentration-dependent cell death with an EC50 of 22.0 μg/mL. PMPMEase activity in the curcumin-treated cell lysate followed a similar concentration-dependent profile with IC50 of 22.6 μg/mL. In colorectal cancer tissue microarray studies, PMPMEase immunoreactivity was significantly higher in 88.6% of cases compared to normal colon tissues (P < 0.0001). The mean scores ± SEM were 91.7 ± 11.4 (normal), 75.0 ± 14.4 (normal adjacent), 294.8 ± 7.8 (adenocarcinoma), and 310.0 ± 22.6 (mucinous adenocarcinoma), respectively. PMPMEase overexpression in colorectal cancer and cancer cell death stemming from its inhibition is an indication of its possible role in cancer progression and a target for chemopreventive agents.

Risk factors and causes of death in MEN1 disease. A GTE (Groupe d'Etude des Tumeurs Endocrines) cohort study among 758 patients

BACKGROUND

The natural history of multiple endocrine neoplasia type 1 (MEN1) is known through single-institution or single-family studies. We aimed to analyze the risk factors and causes of death in a large cohort of MEN1 patients.

METHODS

Overall, 758 symptomatic MEN1 patients were identified through the GTE network (Groupe d'étude des Tumeurs Endocrines), which involves French and Belgian genetics laboratories responsible for MEN1 diagnosis and 80 clinical reference centers. The causes of death were analyzed. A frailty model, including time-dependent variables, was used to assess the impact of each clinical lesion, except for hyperparathyroidism, on survival.

RESULTS

The median follow-up was 6.3 years. Female gender, family history of MEN1, and recent diagnosis were associated with a lower risk of death. Compared with nonaffected patients, those with thymic tumors (hazard ratio [HR] = 4.64, 95% CI = 1.73-12.41), glucagonomas-vipomas-somatostatinomas (HR = 4.29, 95% CI = 1.54-11.93), nonfunctioning pancreatic tumors (HR = 3.43, 95% CI = 1.71-6.88), and gastrinoma (HR = 1.89, 95% CI = 1.09-3.25) had a higher risk of death after adjustment for age, gender, and diagnosis period. The increased risk of death among patients with adrenal tumors was not significant, but three patients died from aggressive adrenal tumors. Pituitary tumors, insulinomas, and bronchial tumors did not increase the risk of death. The proportion of MEN1-related deaths decreased from 76.8 to 71.4% after 1990.

CONCLUSIONS

The prognosis of MEN1 disease has improved since 1980. Thymic tumors and duodenopancreatic tumors, including nonsecreting pancreatic tumors, increased the risk of death. Rare but aggressive adrenal tumors may also cause death. Most deaths were related to MEN1. New recommendations on abdominal and thoracic imaging are required.

Analysis of multiple exposures: an empirical comparison of results from conventional and semi-bayes modeling strategies

BACKGROUND

Analysts of epidemiologic data often contend with the problem of estimating the independent effects of many correlated exposures. General approaches include assessing each exposure separately, adjusting for some subset of other exposures, or assessing all exposures simultaneously in a single model such as semi-Bayes modeling. The optimal strategy remains uncertain, and it is unclear to what extent different reasonable approaches influence findings. We provide an empirical comparison of results from several modeling strategies.

METHODS

In an occupational case-control study of lung cancer with 184 exposure substances, we implemented 6 modeling strategies to estimate odds ratios for each exposure-cancer association. These included one-exposure-at-a-time models with various confounder selection criteria (such as a priori selection or a change-in-the-estimate criterion) and semi-Bayes models, one version of which integrated information on previous evidence and chemical properties.

RESULTS

While distributions of odds ratios were broadly similar across the 6 analytic strategies, there were some differences in point estimates and in substances manifesting statistically significant odds ratios, particularly between strategies with few or no occupational covariates and those with many. Semi-Bayes models produced fewer statistically significant odds ratios than other methods. A simple semi-Bayes model that shrank all the 184 estimates to a common mean yielded nearly identical results to one that integrated considerable prior information.

CONCLUSION

Different modeling strategies can lead to different results. Considering the conceptual and pragmatic difficulties of identifying confounders, these results suggest that it would be unwise to place uncritical reliance on any single strategy.

The impact of time-dependent bias in proportional hazards modelling

In the clinical literature, time-dependent exposure status has regularly been analysed as if known at time origin. Although statisticians agree that such an analysis yields biased results when analysing the effect on the time until some endpoint of interest, this paper is the first to study in detail the bias arising in a proportional hazards analysis. We show that the biased hazard ratio estimate will always be less than the unbiased one; this leads to either an inflated or a damped effect of exposure, depending on the sign of the correct log hazard ratio estimate. We find an explicit formula of the asymptotic bias based on generalized rank estimators, and we investigate the role of censoring, which may prevent an individual from being considered as being baseline exposed in the biased analysis. We illustrate our results with data on hospital infection status and different censoring patterns.

Markov modelling of HIV infection evolution in the HAART era

The aim was to investigate the impact of the main prognostic factors on HIV evolution. A multi-state Markov model was applied in a cohort of 2126 patients to estimate impact of these factors on patients' clinical and immunological evolutions. Clinical progression and immunological deterioration shared most of their prognostic factors: male gender, intravenous drug use, weight loss, low haemoglobin level (<110 g/l), CD8 cell count (<500/mm(3)) and HIV viral load (>5 log(10) copies/ml). Highly active retroviral therapy reduced the risks of clinical progression and immune deterioration whatever patients' CD4 cell count. Risk reductions were 41-60% for protease inhibitor-based and 27-68% for non-nucleoside reverse transcriptase inhibitor-based regimens. Three-year transition probabilities showed that only patients with a CD4 cell count >or=350 CD4/mm(3) could in most cases maintain their immunity. This model provides 'real life' transition probabilities from one immunological stage to another, allowing decision analyses that could help determine the beneficial therapeutic strategies for HIV-infected patients.

Restrictive access to clopidogrel and mortality following coronary stent implantation

BACKGROUND

In Canada, access to clopidogrel is restricted by most provincial drug insurance plans in order to contain costs. Until April 2007, the Régie de l'assurance maladie du Québec (RAMQ) Prescription Drug Insurance Plan reviewed special access forms before approving reimbursement for clopidogrel prescriptions. We investigated the impact of this restrictive process on patient's filling of prescriptions and on all-cause mortality following coronary stenting.

METHODS

We analyzed prescriptions filled and all-cause mortality in the year following a percutaneous coronary intervention among patients who underwent stent implantation between January 2000 and September 2004. We obtained administrative data from the RAMQ databases. We included patients who filled at least 1 prescription for a nonrestricted cardiovascular drug after hospital discharge. We used Cox proportional models to compare mortality rates as a function of delayed or absent outpatient clopidogrel therapy.

RESULTS

Of 13,663 patients, 1571 (11.5%) did not fill any clopidogrel prescription despite filling at least 1 nonrestricted cardiovascular drug prescription after a percutaneous coronary intervention, and 1174 (8.6%) patients filled their clopidogrel prescription with a delay of at least 1 day (median delay 5 days) after filling the nonrestricted cardiovascular drug prescription. After controlling for pertinent covariables, not filling a clopidogrel prescription (hazard ratio [HR] 1.70, 95% confidence interval [CI] 1.35-2.15) and filling with a delay (HR 1.34, 95% CI 1.01-1.80) were associated with a significant increase in all-cause mortality.

INTERPRETATION

Restricted access to clopidogrel was associated with about 20% of patients either not receiving clopidogrel or receiving therapy after a delay. Delay or absence of clopidogrel therapy increased the risk of all-cause mortality after percutaneous coronary intervention with stenting.

tdc.msm: an R library for the analysis of multi-state survival data

The aim of this paper is to present an R library, called tdc.msm, developed to analyze multi-state survival data. In this library, the time-dependent regression model and multi-state models are included as two possible approaches for such data. For the multi-state modelling five different models are considered, allowing the user to choose between Markov and semi-Markov property, as well as to use homogeneous or non-homogeneous models. Specifically, the following multi-state models in continuous time were implemented: Cox Markov model; Cox semi-Markov model; homogeneous Markov model; non-homogeneous piecewise model and non-parametric Markov model. This software can be used to fit multi-state models with one initial state (e.g., illness diagnosis), a finite number of intermediate states, representing, for example, a change of treatment, and one absorbing state corresponding to a terminal event of interest. Graphical output includes survival estimates, transition probabilities estimates and smooth log hazard for continuous covariates.

The Prognostic Value of Initial Relapses on the Evolution of Disability in Patients with Relapsing-Remitting Multiple Sclerosis

The evolution of multiple sclerosis (MS) and the resulting disability are unpredictable. To identify clinical variables that could be potential prognostic factors, we followed a cohort of 288 patients diagnosed as having relapsing-remitting MS between 1990 and 2003. The end point was the first occurrence of a non-reversible EDMUS-GS score >or=3 (moderate disability). The impact of the number of MS attacks during the first 2 years of the disease as well as the first interattack interval were assessed in two Cox models, one using a fixed-in-time covariate, the other using a time-dependent covariate. Older age at onset and a higher number of MS attacks during the first 2 years of MS proved to be predictors of unfavourable prognosis. The first interattack interval had no influence on the evolution of the disability, conversely to the first relapse which had a short-term impact on the prognosis. We confirmed that the age at onset and the number of MS attacks during the first 2 years of MS are predictors of the evolution of the disability and demonstrated the importance of using time-dependent covariates.

A population-based assessment of the prognostic value of the CD19 positive lymphocyte count in B-cell chronic lymphocytic leukemia using Cox and Markov models

No population-based study has assessed the prognostic impact on survival of the CD19 positive lymphocyte count, evaluated by immunophenotyping at diagnosis, in B-cell chronic lymphocytic leukemia (B-CLL). Aiming at addressing this issue, we investigated the clinical outcome of a well-defined population of B-CLL patients. Survival of B-CLL patients, diagnosed between 1990 and 1999 and recorded by the Registry of Hematological Malignancies of the Côte d'Or, was analysed applying Cox's regression model to the 237 included cases and to the 195 Binet stage A patients. To assess simultaneously the predictive value of each parameter on the risk of disease progression and on the risk of death, we completed this analysis by applying a three-states homogeneous Markov model to the whole study population. Analysis of the entire population showed that age (p < 0.001), Binet stage (p = 0.008) and CD19 positive lymphocyte count (p = 0.038) were three independent prognostic factors. However, in stage A patients, only progression into a more advanced stage, analysed as a time-dependent variable, and age had a clear impact on survival (p < 0.001 for both). Markov model revealed that an increased CD19 positive lymphocyte count increased the risk of disease progression in stage A patients (p = 0.002) but did not have direct impact on survival of either stage A patients with stable disease or stage B or C patients. An increased CD19 positive lymphocyte count at diagnosis is a marker of an increased risk of disease progression in stage A patients. Thus, it can be a useful tool for the clinical management of these patients.

Survival bias associated with time-to-treatment initiation in drug effectiveness evaluation: a comparison of methods

The authors compared five methods of studying survival bias associated with time-to-treatment initiation in a drug effectiveness study using medical administrative databases (1996-2002) from Quebec, Canada. The first two methods illustrated how survival bias could be introduced. Three additional methods were considered to control for this bias. Methods were compared in the context of evaluating statins for secondary prevention in elderly patients post-acute myocardial infarction who initiated statins within 90 days after discharge and those who did not. Method 1 that classified patients into users and nonusers at discharge resulted in an overestimation of the benefit (38% relative risk reduction at 1 year). In method 2, following users from the time of the first prescription and nonusers from a randomly selected time between 0 and 90 days attenuated the effect toward the null (10% relative risk reduction). Method 3 controlled for survival bias by following patients from the end of the 90-day time window; however, it suffered a major loss of statistical efficiency and precision. Method 4 matched prescription time distribution between users and nonusers at cohort entry. Method 5 used a time-dependent variable for treatment initiation. Methods 4 and 5 better controlled for survival bias and yielded similar results, suggesting a 20% risk reduction of recurrent myocardial infarction or death events.

Analysis of time-dependent covariates in a regressive relative survival model

Relative survival is a method for assessing prognostic factors for disease-specific mortality. However, most relative survival models assume that the effect of covariate on disease-specific mortality is fixed-in-time, which may not hold in some studies and requires adapted modelling. We propose an extension of the Esteve et al. regressive relative survival model that uses the counting process approach to accommodate time-dependent effect of a predictor's on disease-specific mortality. This approach had shown its robustness, and the properties of the counting process give a simple and attractive computational solution to model time-dependent covariates. Our approach is illustrated with the data from the Stanford Heart Transplant Study and with data from a hospital-based study on invasive breast cancer. Advantages of modelling time-dependent covariates in relative survival analysis are discussed.