Proportional Subdistribution Hazards Regression Research Articles

Sex-based Differences in Risk for Therapy-related Myeloid Neoplasms.

Therapy-related myeloid neoplasm (t-MN) is a life-threatening complication of autologous peripheral blood stem cell (PBSC) transplantation for non-Hodgkin lymphoma (NHL). Prior studies report an association between clonal hematopoiesis (CH) in PBSC and risk of t-MN, but small samples precluded examination of risk within specific sub-populations. Targeted DNA sequencing was performed to identify CH mutations in PBSC from a retrospective cohort of 984 NHL patients (median age at transplant 57y; range: 18-78). Fine-Gray proportional subdistribution hazard regression models estimated association between number of CH mutations and t-MN, adjusting for demographic, clinical, and therapeutic variables. Secondary analyses evaluated association between CH and t-MN among males and females. CH was identified in PBSC from 366 patients (37.2%). t-MN developed in 60 patients after median follow-up of 5y. Presence of ≥2 mutations conferred increased t-MN risk (adjusted hazard ratio [aHR]=2.10, 95% confidence interval [CI]=1.08-4.11, p=0.029). CH was associated with increased t-MN risk among males (aHR=1.83, 95%CI=1.01-3.31) but not females (aHR=0.56, 95%CI=0.15-2.09). Although prevalence and type of CH mutations in PBSC was comparable, the 8y cumulative incidence of t-MN was higher among males vs. females with CH (12.4% vs. 3.6%) but was similar between males and females without CH (4.9% vs. 3.9%). Expansion of CH clones from PBSC to t-MN was seen only among males. Presence of CH mutations in PBSC confers increased risk of t-MN after autologous transplantation in male but not female patients with NHL. Factors underlying sex-based differences in risk of CH progression to t-MN merit further investigation.

Incidence and risk factors for central nervous system (CNS) metastasis (met) in patients (pts) with stage III melanoma treated with adjuvant (ADJ) anti-PD1 immunotherapy (IMT).

9581 Background: The identification of risk factors for CNS met in stage III melanoma pts will facilitate the development of personalized surveillance strategies. Previously we characterized the incidence (3.6%, 9.6%, and 15.8% at 1, 2, and 5 yrs) and risk factors for CNS met in pts diagnosed with stage III melanoma from 1998 to 2014. It is unknown if the incidence of CNS met has changed in the ADJ IMT era. Methods: We performed an institutionally approved retrospective chart review of pts diagnosed with stage III melanoma treated with ADJ IMT between 1/1/2018 and 12/31/2020 at MD Anderson Cancer Center (MDA) or the Melanoma Institute of Australia (MIA). Exclusion criteria included mucosal or uveal melanoma or prior systemic therapy (other than ADJ interferon). Pts had to have CNS imaging within 1 month before or 4 months after stage III diagnosis. Pt and tumor features, treatments, and distant recurrence events (CNS met; non-CNS met) were captured. Time-to-CNS and time-to-non-CNS met were computed from start date of ADJ IMT to date of CNS/non-CNS met. Cumulative incidence of distant met events was determined using competing risks (death) and pts alive with no met at last follow-up were censored. Associations were determined using proportional sub-distribution hazards regression models; group differences were assessed using Gray’s test. Results: 361 pts were included in the analysis (MDA, 212; MIA, 149). With a median FU of 34.1 months (range 1.7-61.9), 128 (35.5%) of pts were diagnosed with distant mets. Among pts with distant mets, 56 had CNS met (15.5% of full cohort; 43.8% of pts with distant met); 24 (18.8% of pts with distant mets) pts had CNS met at initial diagnosis of stage IV disease. The 1-, 2-, and 5-yr rates of CNS met in the full cohort were 7%, 12%, and 17%. There was no difference in CNS met incidence between MDA and MIA. On univariate analysis, risk of CNS met was associated with AJCC stage IIID (Hazard Ratio [HR] 6.4; 95% Confidence Interval [CI] 2.2 – 18.8, p<0.001); increased primary tumor Breslow thickness (HR 1.1, CI 1.1 – 1.2, p<0.001), ulceration (HR 2.8; CI 1.5 – 5.1, p<0.001), and mitotic rate > 9/mm (HR 2.3; CI 1.1 – 4.6, p=0.025). There was a trend for decreased risk of CNS met in pts that stopped ADJ IMT due to toxicity (HR 0.4; CI 0.2 – 1.1, p=0.08). Multivariable (MV) model with these features did not identify any significant (p<0.05) associations with CNS met. Primary tumor ulceration and mitotic rate were significantly associated with non-CNS met on MV analysis. Conclusions: Despite the established impact of ADJ IMT on distant met free survival, the incidence of CNS met in stage III melanoma pts treated with ADJ IMT remains high. The results support the need for continued investigation of risk factors, prevention strategies and optimal surveillance for CNS met in stage III melanoma pts.

Changes in physical activity during the year after the Great East Japan Earthquake and future frailty in older survivors.

This study examines whether changes in physical activity (PA) during the first year after the Great East Japan Earthquake and Tsunami (2011-2012) contributed to preventing the onset of future frailty among older survivors of the disaster. This study tracked 2561 physically active Japanese survivors aged ≥ 65 years (43.6% men; mean age 72.9 years) who had completed self-administered questionnaires in 2011 and 2012. PA levels for participants were classified into four categories based on ≥23 and <23 metabolic equivalent hours/week in 2011 and 2012: "consistently low," "decreasing," "increasing," and "consistently high." Frailty was defined as a Kihon Checklist score ≥ 5, which is used in the long-term care insurance system in Japan. Hazard ratios were calculated for the onset of frailty using a Cox proportional hazards model that fitted the proportional sub-distribution hazards regression model with weights for competing risks of death. From 2012 to 2018, 283 men and 490 women developed frailty. Men with consistently high or increasing PA during the first year after the disaster had a lower risk of frailty. Furthermore, even increasing PA by walking for just 30 min/day prevented future frailty in men; however, this association between a change in PA and the decreased risk of frailty was not observed in women. Older men who remained physically active or resumed PA at an early stage and at a low intensity, even after being physically inactive owing to the disaster, were able to prevent future frailty. Geriatr Gerontol Int 2024; 24: 563-570.

Comparison of Infection Rates between Single-Lumen and Double-Lumen Chest Ports among Patients with Cancer: A Propensity Score Matching Analysis

PurposeTo compare the port infection rate between single-lumen (SL) and double-lumen (DL) ports and to determine whether the use of a DL port is an independent risk factor for port infection among patients with cancer. Materials and MethodsThis retrospective study included 2,573 adult oncologic patients (aged >18 years) who had either a SL (n = 841) or a DL (n = 1,732) chest port implanted between 2013 and 2020 at a single institution. Patients who had port infection, including port-site infection and port-related bloodstream infection, were identified through chart review. After propensity score matching based on 13 potentially confounding variables, a total of 493 pairs of patients with either SL (SL group) or DL (DL group) ports were subjected to analysis. The port infection rate was compared between the 2 groups using Poisson regression. Multivariate proportional subdistribution hazards regression (PSHREG) analysis was conducted to determine whether use of a DL port is an independent risk factor for port infection. ResultsThe cumulative follow-up period for the matched cohort was 371,853 catheter-days (median, 297 catheter-days per port; range, 0–1,903 catheter-days). The port infection rate of the DL group was significantly higher than that of the SL group (0.232 vs 0.113 infections per 1,000 catheter-days; P = .001). PSHREG analysis demonstrated that use of a DL port was an independent risk factor of port infection (subdistribution hazard ratio, 2.30; 95% CI, 1.33–3.78; P = .002). ConclusionsDL ports were associated with a higher risk of port infection compared with SL ports in adult oncologic patients.

Association between High Antimetabolite Dose Intensity during Maintenance, Toxicities and Relapse Risk in Children with Acute Lymphoblastic Leukemia (ALL): COG-AALL03N1 Report

Introduction: Current COG protocols for childhood ALL recommend increasing 6-mercaptopurine (6MP) and methotrexate (MTX) dose intensity (DI=prescribed therapy dose/ planned protocol dose) to &amp;gt;100% during maintenance for patients exhibiting inadequate myelosuppression (surrogate for systemic exposure to 6MP/MTX). Whether increasing DI &amp;gt;100% (high DI) to achieve myelosuppression is associated with increased toxicity is not clear. Additionally, whether high DI influences relapse risk, and the direction of that association, remains unknown. Nonadherence to 6MP results in inadequate myelosuppression and prompts clinicians to increase the prescribed dose. Nonadherence to 6MP is also associated with relapse (Bhatia, JAMA Onc 2015). Prior literature examining the association between high DI and relapse risk did not consider adherence to chemotherapy. Here, we examined whether high 6MP/MTX DI affects toxicity or relapse, taking 6MP adherence into consideration. Methods: We used data collected by COG-AALL03N1 (primary aim: 6MP adherence and risk of relapse) to examine the association between high DI and (i) hematologic toxicity, and (ii) relapse risk. Inclusion criteria: (i) age &amp;lt;21y at ALL diagnosis, (ii) receiving maintenance therapy in first remission and &amp;gt;6mo since starting maintenance, (iii) wild-type thiopurine methyltransferase ( TPMT) or nucleoside diphosphate-linked motif 15 (NUDT15) and (iv) objectively measured 6MP adherence for 6mo per patient (using MEMS-TrackCaps). 6MPDI and MTXDI were calculated monthly for 6mo. Updates on vital status, relapse or second neoplasm were collected every 6mo-1y from participating sites for up to 10y. Exposure: ‘high DI’ phenotype: (i) 6MPDI or MTXDI ≥110% at any time during the first 4 study months, or (ii) 6MPDI or MTXDI 100% to 110% at study enrollment and ≥25% increase over the first 4 study months. All other patients were categorized as having ‘normal DI’. Outcomes: (i) hematologic toxicity (absolute neutrophil count &amp;lt;500 cells/µL, platelet count &amp;lt;50,000 cells/mm 3, or treatment interruption due to low ANC/platelets) during the subsequent 2 study months (months 5 and 6); and (ii) ALL relapse any time after month 4. We used logistic regression to examine the association between high DI and hematologic toxicity, adjusting for age at study entry, sex, race/ethnicity. The association between DI and relapse was estimated using proportional sub-distribution hazard regression models, adjusting for age at study entry, blast cytogenetics, time from start of maintenance, and treatment interruptions. Analyses were stratified by 6MP adherence (adherers: median adherence ≥85%; non-adherers: median adherence &amp;lt;85%). Adherence was included in both models to account for residual confounding. Results: Characteristics of the 410 eligible participants are summarized in the Table; the cohort was followed for a median 7.9y (range, 0.2-13). There were no significant differences between patients with normal (71.5%) vs high (28.5%) DI. Patients with high DI had lower median 6MP adherence (94.6% vs 97.0%, P=0.02). Hematologic DLT: As shown in the Figure, the proportion of patients with hematologic toxicities was significantly higher among adherers with high DI compared with adherers with normal DI (38.1% vs 18.5%, P=0.0003). In contrast, the proportion of patients with hematologic toxicities was not significantly different by DI status among non-adherers (18.5% vs 10.5%, P=0.4). Multivariable analysis revealed that high DI was associated with 2.9-fold greater odds of hematologic toxicity among adherers (95%CI=1.6-5.1, P&amp;lt;0.001) and a non-significant 2.1-fold greater odds of hematologic toxicity among non-adherers (95%CI=0.4-10.1, P=0.4). Relapse risk: The adjusted hazard of relapse among adherent patients exposed to high DI was 2.2-fold greater (95%CI=1.0-5.1, P=0.06) than adherent patients exposed to normal DI after adjusting for covariates. In contrast, the hazard of relapse among non-adherent patients exposed to high DI (vs. normal DI) was not elevated (HR=0.9, 95%CI=0.2-3.8, P=0.8). Conclusions: High antimetabolite dose intensity in children with ALL who adhere to oral 6MP during maintenance therapy is associated with higher odds of hematologic toxicities as well as increased hazard of relapse. These findings inform clinicians about the issues related to dose escalation beyond protocol doses, especially among adherent patients.

Rapid weight change as a predictor of disability among community-dwelling Japanese older adults.

To fill the knowledge gap regarding weight change and the onset of disability in community-dwelling Japanese older adults, we investigated the potential effects of rapid weight change on disability risk as defined by Japan's long-term care insurance (LTCI) system. We analyzed data from a longitudinal study of 10 375 community-dwelling older Japanese adults (≥65 years) who were not LTCI needs certified at baseline and joined the study from 2002 to 2005. Weight change (percentage) was calculated by subtracting participants' weight in the previous year from that measured during a physical examination at study commencement. The five weight-change categories ranged from sizable weight loss (≤ -8.0%) to sizable weight gain (≥ +8.0%). Disability was defined according to LTCI certifications at follow-up. Hazard ratios (HRs) and 95% confidence intervals were calculated for new-onset disability using a Cox proportional hazards model that fitted the proportional subdistribution hazards regression model with weights for competing risks of death. During the mean 10.5-year follow-up, 2994 participants developed a disability. Sizable weight loss (HR [95% confidence intervals], 1.41 [1.17-1.71]) and weight loss (1.20 [1.05-1.36]) were significant predictors of disability onset. Sizable weight gain (1.45 [1.07-1.97]) corresponded to severe disability. Stratified analyses by lifestyle and initial body mass index categories revealed more pronounced associations between weight change and disability risk in the unhealthy lifestyle and below initial normal body mass index groups. Rapid and sizable weight gain could be additional criteria for disability risk in older adults. Geriatr Gerontol Int 2023; 23: 809-816.

Abstract No. 20 Right- versus Left-Sided Chest Ports in Oncologic Patients: Are There Any Differences in Complication Rates after Propensity Score Matching?

There is a lack of consensus as to whether the laterality of central venous chest port (port) placement is a risk factor for port complications. This is partly because confounding variables were suboptimally controlled in prior studies. The purpose of this study is to retrospectively compare the port-related complication rate between right-sided and left-sided ports placed in oncologic patients using propensity score matching. A total of 3,126 consecutive adult (≥18 years old) oncologic patients (male/female: 1,571/1,555, median age: 62) who had a port placed between 2013 and 2021 were included in this study. Patients with port complications (infectious, mechanical, thrombotic, and skin complications) were identified through chart review. Propensity score matching generated a cohort of 277 pairs of patients for the right-sided and left-sided port groups that are balanced across 11 potentially confounding variables. The overall complication rate of each group was compared using Poisson regression. Multivariate proportional subdistribution hazard regression (PSHREG) analysis was also conducted to determine if port laterality is an independent risk factor for overall port complications. The cumulative follow-up period for the matched cohorts was 259,250 catheter-days (median per patient: 311 catheter-days, range:10–2,879 catheter-days). In the left-sided port group, a total of 59 ports (21.3%) were found to be associated with complications (36 infectious, 7 mechanical, 11 thrombotic, 12 skin-related), while 31 ports in the right-side port group (11.2%, P = 0.0013) were associated with complications (13 infections, 3 mechanical, 8 thrombotic, 8 skin-related), accounting for complication rates of 0.46 and 0.24 complications/1000 catheter-days, respectively (P = 0.0022). PSHREG analysis of the matched cohort further demonstrated laterality of port as an independent risk factor for port complication (hazard ratio = 2.015, 95% confidence interval = 1.306-3.108, P = 0.0015). Left-sided ports were associated with a higher risk of port complications when compared with right-sided ports in adult oncological patients based on propensity score matching analysis. The right-sided port placement is recommended to minimize the risk of port complications.

Assessment of a serum calcification propensity test for the prediction of all-cause mortality among hemodialysis patients

BackgroundVascular calcification is a major contributor to the high cardiac burden among hemodialysis patients. A novel in vitro T50-test, which determines calcification propensity of human serum, may identify patients at high risk for cardiovascular (CV) disease and mortality. We evaluated whether T50 predicts mortality and hospitalizations among an unselected cohort of hemodialysis patients.MethodsThis prospective clinical study included 776 incident and prevalent hemodialysis patients from 8 dialysis centers in Spain. T50 and fetuin-A were determined at Calciscon AG, all other clinical data were retrieved from the European Clinical Database. After their baseline T50 measurement, patients were followed for two years for the occurrence of all-cause mortality, CV-related mortality, all-cause and CV-related hospitalizations. Outcome assessment was performed with proportional subdistribution hazards regression modelling.ResultsPatients who died during follow-up had a significantly lower T50 at baseline as compared to those who survived (269.6 vs. 287.7 min, p = 0.001). A cross-validated model (mean c statistic: 0.5767) identified T50 as a linear predictor of all-cause-mortality (subdistribution hazard ratio (per min): 0.9957, 95% CI [0.9933;0.9981]). T50 remained significant after inclusion of known predictors. There was no evidence for prediction of CV-related outcomes, but for all-cause hospitalizations (mean c statistic: 0.5284).ConclusionT50 was identified as an independent predictor of all-cause mortality among an unselected cohort of hemodialysis patients. However, the additional predictive value of T50 added to known mortality predictors was limited. Future studies are needed to assess the predictive value of T50 for CV-related events in unselected hemodialysis patients.

Gene Screening for Prognosis of Non-Muscle-Invasive Bladder Carcinoma under Competing Risks Endpoints.

Background: Discovering clinically useful molecular markers for predicting the survival of patients diagnosed with non−muscle-invasive bladder cancer can provide insights into cancer dynamics and improve treatment outcomes. However, the presence of competing risks (CR) endpoints complicates the estimation and inferential framework. There is also a lack of statistical analysis tools and software for coping with the high-throughput nature of these data, in terms of marker screening and selection. Aims: To propose a gene screening procedure for proportional subdistribution hazards regression under a CR framework, and illustrate its application in using molecular profiling to predict survival for non-muscle invasive bladder carcinoma. Methods: Tumors from 300 patients diagnosed with bladder cancer were analyzed for genomic abnormalities while controlling for clinically important covariates. Genes with expression patterns that were associated with survival were identified through a screening procedure based on proportional subdistribution hazards regression. A molecular predictor of risk was constructed and examined for prediction accuracy. Results: A six-gene signature was found to be a significant predictor associated with survival of non−muscle-invasive bladder cancer, subject to competing risks after adjusting for age, gender, reevaluated WHO grade, stage and BCG/MMC treatment (p-value < 0.001). Conclusion: The proposed gene screening procedure can be used to discover molecular determinants of survival for non−muscle-invasive bladder cancer and in general facilitate high-throughput competing risks data analysis with easy implementation.

Specific Clonal Mutations Are Predictive of Therapy-Related Myeloid Neoplasms (tMN) after Autologous Peripheral Blood Stem Cell Transplantation (aPBSCT) for Lymphoma

Background: tMN is a lethal complication of aPBSCT for Hodgkin (HL) and non-Hodgkin lymphoma (NHL). Clonal somatic mutations in leukemia-associated genes (clonal hematopoiesis [CH]) are observed in peripheral blood of healthy older adults, and are associated with an elevated risk for subsequent leukemia. In cancer patients, CH mutations are differentially selected by therapeutic exposures. We examined the association between CH in the PBSC product and risk of post-aPBSCT tMN. Methods: We constructed a retrospective cohort of 1,328 patients (NHL=1,000; HL=328) who received aPBSCT between 1980 and 2016 at a single institution. Information on demographics (sex, race), clinical characteristics (HL/NHL, age at aPBSCT, CD34 counts in PBSC product) and therapeutic exposures (pre-aPBSCT cyclophosphamide equivalent dose [CED], doxorubicin equivalence dose, radiation; total body irradiation [TBI]) was abstracted from medical records. Using cryopreserved aliquots of PBSC product from all patients, we performed sequencing of target regions in 91 genes selected for association with CH or myeloid malignancies (Illumina NovaSeq platform; >1000x), to identify pathogenic mutations (variant allele frequency [VAF] ≥1%). Logistic regression identified demographic/clinical/treatment factors associated with CH. Fine-Gray proportional sub-distribution hazard regression determined the role of CH (yes/no; number of mutations [0, 1, 2+]; specific mutations) in development of subsequent tMN, adjusting for above factors. We also examined the role of CH in post-aPBSCT mortality. Results: Median age at aPBSCT was 50y (range: 18-78); 62% male; 61% non-Hispanic white. CH in the PBSC was identified in 435 patients (33%); single mutation: 276 (21%); >1 mutation: 159 (12%). Most commonly mutated genes included DNMT3A (192, 14.5%), PPM1D (118, 8.9%), TET2 (82, 6.2%), TP53 (44, 3.3%). Seventy-five patients developed tMN, yielding an 8y cumulative incidence of 5.9%. Predictors for CH in PBSC: Older age at aPBSCT (odds ratio [OR]per_y_inc=1.07, p=5.9x10-34 and high pre-aPBSCT alkylating agent exposure [CED=1-12,000: OR=1.58, p=0.01; CED >12,000: OR=2.03, p=0.057) were independently associated with CH. CH in PBSC and post-aPBSCT t-MN: Among the 435 patients with CH, 38 (9%) developed tMN. The 8y cumulative incidence of tMN was highest in patients with >1 mutations (14.6%) vs. single mutation (6.9%), vs. no mutation (4.0%), p=2.4x10-6. Multivariable analysis (Table 1) revealed >1 mutation as an independent predictor of tMN (HR=2.38, p=0.003; single mutation: HR=1.27, p=0.47; reference: no mutations); other factors associated with tMN are in Table 1. Specific mutations in PBSC andpost-aPBSCTtMN: The role of specific mutations in the development of tMN is detailed in Table 2. tMN incidence in patients with DNMT3A mutations alone did not differ from patients with no mutations (8y incidence: 4.2% vs. 4.0%, p=0.95), but was increased in patients with PPM1D (9.4% vs. 4.0%, p=0.06) and TP53 (9.5% vs. 4.0%, p=0.22) mutations alone, although not reaching statistical significance. The co-occurrence of mutations in DNMT3A + PPM1D (21.4% vs. 4.0%, p=0.0006), DNMT3A +TP53 (33.3% vs. 4.0%, p <0.001) or PPM1D + TP53 (42.9% vs. 4.0%, p <0.001) was associated with significantly higher tMN incidence compared with no CH mutations. Taken together, individuals with the combination of mutations in genes involved in epigenetic regulation (DNM3TA, TET2 and/or ASXL1) and DNA repair (TP53, PPM1D, ATM, and/or SRCAP) had significantly higher tMN incidence compared to those with mutations in genes involved in epigenetic regulation alone (26.6% vs. 5.4%, p=1.1x10-5). Specific mutations at time of tMN: In available samples (peripheral blood or bone marrow: n=40), TP53 (n=13, 32.5%), PPM1D (n=12, 30%), DNMT3A (n=12, 30%) and TET2 (n=4, 10%) were the most frequently mutated genes. CH in PBSC and mortality: All-cause mortality at 8y from aPBSCT was significantly higher in patients with CH (38.4%) vs. those without (30.1%), p=0.0009. Conclusion and Relevance: CH is associated with increased risk of tMN after aPBSCT for lymphoma, but most patients with CH do not develop tMN. This study shows that the co-occurrence of multiple key somatic mutations in the PBSC product increases tMN risk. Importantly, mutations in DNA repair/ damage response genes, especially in combination with epigenetic regulatory genes, increase the risk of tMN. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Poverty during Maintenance and Relapse Risk in Children with Acute Lymphoblastic Leukemia (ALL): A Report from Children's Oncology Group Study (COG-AALL03N1)

Introduction: There is conflicting evidence regarding the association between poverty and survival in children with cancer. Most studies have relied on surrogate indicators of poverty, such as zip code-based median income for area of residence or on insurance status (Bona, PBC 2016; Kehm, Cancer 2019; Bona, JNCI 2021). We hypothesized that household poverty (measured at the individual level, using self-reported annual household income) during maintenance would be associated with a greater hazard of relapse among children with ALL. Methods: In this secondary analysis of COG-AALL03N1 (primary aim: adherence to oral mercaptopurine during maintenance), we examined the association between household poverty during maintenance and relapse risk in children with ALL. Eligibility for enrollment on AALL03N1 included (1) age <21y at ALL diagnosis; (2) receiving maintenance therapy in first remission. The current analysis was restricted to patients living in US who provided information on annual household income and the number of household members. We used national yearly median poverty thresholds provided by CDC; these take into account the annual household income and the size of family unit (i.e., number of adults and children <18y). For example, the annual income of a three-person household (two parents and one child [patient]) would need to be >$14,480 to be above the poverty threshold in the year 2002 (or >$21,831 in the year 2021). Among those below poverty threshold, we classified families with annual income under 120% federal threshold (median percent below poverty threshold for this cohort) as living in extreme poverty. Hazard of relapse from study enrollment during ALL maintenance therapy was estimated using proportional subdistribution hazard regression models, treating extreme poverty as the primary exposure and adjusting for age at study, sex, NCI risk group, blast cytogenetics, average 6MP and methotrexate dose intensity and time from start of maintenance to study entry. Results: We summarize sociodemographic and disease characteristics of 592 eligible patients in the Table. Median household income of the cohort was $35,000/year (range $10,000->$150,000). Overall, 73 (12.3%) patients met criteria for extreme poverty. Families living in extreme poverty were more likely to self-report their race/ethnicity as Hispanic or African-American and report parental education as high school (HS) or lower. Of note, clinical factors did not differ by poverty level. In a logistic regression, non-white race/ethnicity (odds ratio [OR]=7.9, 95%CI=2.4-26.1, P=0.007; reference: non-Hispanic white) and low parental education (≤HS: OR=4.2, 95%CI=2.4-7.4, P=<0.001; reference: college degree or higher) were associated with greater odds of living in extreme poverty. Cumulative incidence of relapse at 3y (Figure) from study entry among patients living in extreme poverty was higher compared to those without extreme poverty (14.3% [95%CI, 7-24] vs 7.6% [95%CI, 5-10], P=0.04). Patients living in extreme poverty had a 1.97-fold higher hazard of relapse compared to those not living in extreme poverty after adjusting for above mentioned covariates (95%CI=1.03-3.76, P=0.04). The reasons for the association between individual level poverty and relapse risk are likely multifactorial and possibly include poor adherence to oral chemotherapy. We found that patients living in extreme poverty had lower adherence to 6MP compared to those not living in extreme poverty (85.2±19.6% vs 90.6±14.8%, P=0.03). Conclusions: Extreme household poverty is associated with increased hazard of relapse in children receiving maintenance therapy for ALL compared to those living without extreme household poverty. Mechanisms investigating this association need to be explored in future studies. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Risk of non-colorectal cancer-related death in elderly patients with the disease: A comparison of five preoperative risk assessment indices.

A considerable number of elderly patients with colorectal cancer (CRC) die of non-CRC-related causes. The Controlling Nutritional Status (CONUT) score, American Society of Anesthesiologists Physical Status classification, Charlson Comorbidity Index, National Institute on Aging, and National Cancer Institute Comorbidity Index, and Adult Comorbidity Evaluation-27 score are all known predictors of survival in patients with CRC. However, the utility of these indices for predicting non-CRC-related death in elderly CRC patients is not known. The study population comprised 364 patients aged 80 years or more who received curative resection for stage I-III CRC between 2000 and 2016. The association of each index with non-CRC-related death was compared by competing-risks analysis such as the cumulative incidence function and proportional subdistribution hazards regression analysis as well as time-dependent receiver-operating characteristic (ROC) analysis. There were 85 deaths (40 CRC-related and 45 non-CRC-related) during a median observation period of 53.2months. Cumulative incidence function analysis identified CONUT score as the most suitable for risk stratification for non-CRC-related death. In proportional subdistribution hazards regression, risk of non-CRC-related death increased significantly as CONUT score worsened (2/3/4 vs. 0/1, hazard ratio 1.73, 95% confidence interval [CI] 0.91-3.15; ≥5 vs. 2/3/4, hazard ratio 2.71, 95% CI 1.08-6.81). Time-dependent ROC curve analysis showed that CONUT score were consistently superior to other indices during the 5-year observation period. The majority of deaths in elderly patients with CRC were not CRC-related. CONUT score was the most useful predictor of non-CRC-related death in these patients.

Outcomes of Hypofractionated Stereotactic Radiotherapy for Small and Moderate-Sized Brain Metastases: A Single-Institution Analysis.

BackgroundStereotactic radiosurgery (SRS) is the standard treatment for limited intracranial metastases. With the advent of frameless treatment delivery, fractionated stereotactic radiotherapy (FSRT) has become more commonly implemented given superior control and toxicity rates for larger lesions. We reviewed our institutional experience of FSRT to brain metastases without size restriction.MethodsWe performed a retrospective review of our institutional database of patients treated with FSRT for brain metastases. Clinical and dosimetric details were abstracted. All patients were treated in 3 or 5 fractions using LINAC-based FSRT, did not receive prior cranial radiotherapy, and had at least 6 months of MRI follow-up. Overall survival was estimated using the Kaplan–Meier method. Local failure and radionecrosis cumulative incidence rates were estimated using a competing risks model with death as the competing risk. Univariable and multivariable analyses using Fine and Gray’s proportional subdistribution hazards regression model were performed to determine covariates predictive of local failure and radionecrosis.ResultsWe identified 60 patients and 133 brain metastases treated at our institution from 2016 to 2020. The most common histologies were lung (53%) and melanoma (25%). Most lesions were >1 cm in diameter (84.2%) and did not have previous surgical resection (88%). The median duration of imaging follow-up was 9.8 months. The median survival for the whole cohort was 20.5 months. The local failure at 12 months was 17.8% for all lesions, 22.1% for lesions >1 cm, and 13.7% for lesions ≤1 cm (p = 0.36). The risk of radionecrosis at 12 months was 7.1% for all lesions, 13.2% for lesions >1 cm, and 3.2% for lesions ≤1 cm (p = 0.15).ConclusionsFSRT is safe and effective in the treatment of brain metastases of any size with excellent local control and toxicity outcomes. Prospective evaluation against single-fraction SRS is warranted for all lesion sizes.

CHESS-ALARM score to stratify decompensation risk in compensated advanced chronic liver disease patients: An international multicenter study.

A combination of platelet and elastography (PE criteria) was proposed to identify compensated advanced chronic liver disease (cACLD) patients at risk of liver decompensation. We aim to validate and refine PE criteria by developing a new predictive score to predict decompensation in Asian cACLD patients. An international cohort of 633 cACLD patients with liver stiffness measurement (LSM) and esophagogastroduodenoscopy performed were included. We validated PE criteria to predict first liver decompensation using competing risk analysis, with death and hepatocellular carcinoma as competing events. We developed a predictive model using proportional subdistribution hazard regression. Prognostic accuracy was compared with the model of end-stage liver disease (MELD), albumin-bilirubin (ALBI), and ALBI-FIB-4 score using time-dependent area under operative characteristic curve (tAUC). Sixty patients developed decompensation over the median follow-up of 39months. Favorable Baveno VI status ruled out cACLD patients at risk of liver decompensation. LSM>25kPa was suboptimal to predict cACLD patients who will develop liver decompensation. We developed CHESS-ALARM score by incorporating age, platelet, and gender into LSM. CHESS-ALARM score (tAUC=0.86, 95% confidence interval [CI]: 0.79-0.94) has significantly higher accuracy than MELD (tAUC: 0.61), ALBI (tAUC: 0.62), ALBI-FIB-4 (tAUC: 0.70), and LSM>25kPa (tAUC: 0.54) to predict liver decompensation at 5years (P<0.05 for all). Patients with CHESS-ALARM score≥-0.37 had an 11-fold higher risk of decompensation (subdistribution hazard ratio=11.2, 95% CI: 5.1-24.5). CHESS-ALARM score can be readily incorporated into clinical practice of cACLD patients to estimate individual risk of liver decompensation; however, more data are required in morbidly obese cACLD patients of nonviral etiology.

Competing-risk nomogram for predicting survival in patients with advanced (stage III/IV) gallbladder cancer: A SEER population-based study.

The primary aim of this study was to assess the cumulative incidence of cause-specific mortality (CSM) and other cause-specific mortality (OCSM) for patients with advanced gallbladder cancer (GBC), and then to develop a nomogram based on competing-risk analysis to forecast CSM. We identified the patients with GBC with specific screening criteria and from the Surveillance Epidemiology and End Results (SEER) database. We calculated the cumulative incidence function for CSM and OCSM, and constructed a competing-risk nomogram based on the Fine and Gray's proportional subdistribution hazard regression model to forecast the probability of CSM of these patients. In addition, the concordance index and calibration plot were performed to validate the novel established model. A total of 1411 patients were included in this study. The 1-, 2-, and 3-year overall cumulative mortalities were 46.2, 62.2, and 69.6% for CSM, respectively, while they were 6.2, 8.7, and 10.4% for OCSM. Additionally, the 1-, 2-, and 3-year estimates of overall survival were 47.6, 29.1, and 19.9% for above these patients, respectively. We also developed a competing-risk nomogram to estimate the CSM. The concordance index was 0.775 (95% confidence interval (CI): 0.750-0.800) in the training set and that was 0.765 (95% CI: 0.730-0.800) in the internal validation set, which suggests the robustness of the novel established model. Furthermore, the calibration curves and concordance index demonstrated that the nomogram was well-calibrated and demonstrated good discriminative ability. The ample sample allowed us to develop a reliable model which demonstrated better calibration and discrimination for predicting the probability of CSM of patients with advanced GBC.

Kidney Transplantation Outcome Predictions (KTOP): A Risk Prediction Tool for Kidney Transplants from Brain-dead Deceased Donors Based on a Large European Cohort

BackgroundEuropean kidney donation shortages mandate efficient organ allocation by optimizing the prediction of success for individual recipients. ObjectiveTo develop the first European online risk tool for kidney transplant outcomes on the basis of recipient-only and recipient plus donor characteristics. Design, setting, and participantsWe used individual recipient and donor risk factors and three outcomes (death, death with functioning graft [DWFG], and graft loss) for 32 958 transplants within the Eurotransplant kidney allocation system and the Eurotransplant senior program between January 2006 and May 2018 in eight European countries to develop and validate a risk tool. Outcome measurements and statistical analysisCox proportional-hazards models were used to analyze the association of risk factors with overall patient mortality, and proportional subdistribution hazard regression models for their association with graft loss and DWFG. Prediction models were developed with recipient-only and recipient-donor risk factors. Sensitivity analyses based on time-specific area under the receiver operating characteristic curve (AUC) with leave-one-country-out validation were performed and calibration plots were generated. Results and limitationsThe 10-yr cumulative incidence rate was 37% for mortality, 12% for DWFG, and 41% for graft loss. In recipient-donor models the leading risk factors for mortality were recipient diabetes (hazard ratio [HR] 10.73), retransplantation (HR 3.08 per transplant), and recipient age (HR 1.08). Effects were similar for DWFG. For graft loss, diabetes (subdistributional HR [SHR] 1.32), increased donor age (SHR 1.02), and prolonged cold ischemia time (SHR 1.02) had increased SHRs. All p values were <0.001. ConclusionsPreviously identified risk factors for outcomes following kidney transplants allow for outcome prediction with 10-yr AUC values of up to 0.81. Patient summaryUsing European data, we estimated individual risks to predict the success of kidney transplants and support physicians in decision-making. An online tool is now available (https://riskcalc.org/ktop/) for predicting kidney transplant outcomes both before and after a donor has been identified.

Outcome of revised metal-on-metal hip arthroplasties: a Dutch arthroplasty register study.

Preliminary results of metal-on-metal (MoM) hip arthroplasty were satisfactory, but since 2004 data showed high failure rates. National joint replacement registries are multi-centre databases comprised of thousands of subjects and implants which allow for identifying variables predictive of implant failure. The aim of the current study was to estimate re-revision rates after revision of a primary MoM hip arthroplasty in the Dutch Arthroplasty Register (LROI) and to assess potential predictor variables of re-revision of these MoM hip arthroplasties. Eligible procedures were those with a revision for any reason except infection, after an initial primary surgery with a hip resurfacing (HRA) or large-head MoM (LH-MoM) total hip arthroplasty (THA). The probability of re-revision for both types of MoM hip arthroplasty over time was estimated using the cumulative incidence function taking mortality as a competing risk into account. A proportional sub-distribution hazards regression model was used to assess potential predictor variables of re-revision of these MoM hip arthroplasties. A total of 3476 records of revised implants were included, of which 873 (25.2%) were MoM implants. Over the course of follow-up, 101 (11.5%) MoM implants were re-revised. During follow-up 36 (4.3%) patients who received a MoM-implant at primary arthroplasty and a revision afterwards had died. The regression model showed that for primary MoM implants a MoM articulation after revision (HR 2.48; 95% CI 1.53-4.03, p < 0.001), femoral-only revisions (HR 3.20; 95% CI 2.06-4.99, p < 0.001) and periprosthetic fractures (HR 1.98; 95% CI 1.03-3.82, p = 0.042) as reason for the first revision were statistically significant risk factors for re-revision. Both types of large-head MoM hip arthroplasties have shown high revision and re-revision rates; risk factors were identified. The outcome of this study can be helpful in managing expectations of patients and orthopaedic surgeons.

Differential Second Primary Malignancy Occurrence After Breast Cancer According to HER2 Status: A Population-Based Study

PurposeOur study aims to analyze the association between HER2 status of breast cancer (BC) and second primary malignancy (SPM) occurrence.Materials and MethodsBased on BC patients registered between 2010 and 2018 in the NCI SEER database, we utilized standardized incidence ratio (SIR) and Poisson regression to quantify SPM occurrence compared with the general population. Then, adjusted for competing death risk, cumulative incidence function and Gray’s test were adopted to estimate the probability of SPM. Subsequent proportional subdistribution hazards regression was executed to identify the HER2 status impact on SPM risk. Finally, survival analysis was performed.ResultsA total of 409,796 first BC patients were included and 18,283 were identified with at least one SPM. The SIR of SPM after HER2+ BC was significantly lower than HER2- BC (1.03 vs 1.13; RR, 0.92; 95% CI, 0.88–0.96; p<0.001). The predominantly declining SPM risk was only observed for second BC (RR, 0.89; 95% CI, 0.82–0.96; p=0.003) and lung cancer (RR, 0.84; 95% CI, 0.74–0.95; p=0.007). Furthermore, competing risk analysis verified the protective effect of HER2 positivity status on SPM occurrence. The 5-year cumulative incidence of SPM following HER2+ and HER2- BC were 4.09% and 5.16%, respectively (p<0.001). In addition, among patients suffering from SPM, HER2 positivity status contributed to better overall survival.ConclusionIt is demonstrated that HER2+ BC patients had lower SPM incidence, which was remarkable for second BC and lung cancer.

Nonmetastatic breast cancer patients subsequently developing second primary malignancy: A population-based study.

BackgroundWith life span extending, breast cancer (BC) survivors may face the possibility of developing second primary cancer (SPC) and considerably shorten survivorship. However, little is known about multiple primary cancer (MPC) patients with nonmetastatic breast cancer as a first primary malignancy (BCFPM).MethodsHere, we retrospectively analyzed data on cancer survivors with BCFPM diagnosed between 2010 and 2015 from the Surveillance, Epidemiology, and End Results (SEER) database. The prognostic factors for breast cancer‐specific survival (BCSS) were ascertained by the stepwise regression analysis and a competing risk model, and were integrated to the establishment of prognostic nomogram, of which the accuracy was measured by the calibration curve and the concordance index (C‐index).ResultsIn total, 8616 patients were identified with 4.6% of 3‐year breast cancer‐ specific death (BCSD) and 8.6% of 5‐year BCSD. The most common SPC among BCFPM patients were female BC and lung cancer. Besides, the median latency time between BC and SPC was 22 months. At a ratio of 7:3, all patients were randomly categorized into a training cohort (n = 6032) and a validation cohort (n = 2584). By a proportional subdistribution hazards regression analysis, the following factors were considered to own independent prognostic abilities of BCSS: subtypes, grade, T classification, N classification, radiation, and sites of SPC. The nomogram could accurately predict 3‐year and 5‐year breast cancer‐associated survival of BCFPM patients with high internal and external validated C‐index, 0.715 (95% CI, 0.691–0.739), and 0.683 (95% CI, 0.642–0.724), respectively.ConclusionsBC survivors remained a high risk of developing SPC and considerably shortened survival time. In this study, a favorable nomogram was constructed to as a prediction model for 3‐year and 5‐year BCSS of BCFPM patients, largely intending to prolong the life of these patients by assisting clinicians to make individualized follow‐up plans.

Nonmyeloablative, HLA-Mismatched Unrelated Peripheral Blood Transplantation with High-Dose Post-Transplantation Cyclophosphamide

High-dose post-transplantation cyclophosphamide (PTCy) is an effective platform for prevention of severe graft-versus-host disease (GVHD) after allogeneic bone marrow (BM) transplantation with mismatched unrelated donors (mMUDs). Previous studies evaluating PTCy with mMUDs favored BM allografts over peripheral blood stem cell transplantation (PBSCT) due to concerns that PBSCT may be associated with an increased risk of acute and chronic GVHD. In addition, haploidentical PBSCT is associated with high rates of cytokine release syndrome (CRS), which is another concern with mMUD PBSCT. This study was conducted to determine the feasibility and safety of using mMUD PBSCT with PTCy as GVHD prophylaxis. Patients who received mMUD PBSCT using a PTCy-based GVHD prophylaxis at Johns Hopkins Hospital as part of a prospective clinical trial of mMUD and non-first-degree relative haploidentical transplantation with PTCy (ClinicalTrials.gov identifier NCT01203722) were included. All patients underwent T cell-replete PBSCT between November 2012 and August 2020. Statistical analyses were performed using the Kaplan-Meier method and proportional subdistribution hazard regression model for competing risks. The 29 patients in the study had a median age of 54 years, with 10 patients (34%) age ≥60 years. Nineteen grafts (66%) were matched for 9/10 HLA loci, 6 (21%) were match for 8/10, and 4 (14%) were matched for 7/10. No primary or secondary graft failure occurred. The median time to neutrophil recovery (≥500/µL) was 17 days, and that to platelet recovery (≥20,000/µL) was 28 days. Full donor chimerism was achieved in all patients by day +60. The cumulative incidence (CuI) of grade II-IV acute GVHD at 180 days was 15% (90% confidence interval [CI], 3% to 26%). There were no cases of severe chronic GVHD, 3 cases of mild chronic GVHD, and 1 case of moderate chronic GVHD. The CuI of nonrelapse mortality (NRM) was 7% (90% CI, NA to 18%) at 1 year. Eighteen patients (62%) experienced mild CRS (grade 1-2), and 1 patient (3%) experienced severe CRS (grade 3-5). At 1 year, the CuI of relapse was 29% (90% CI, 8% to 50%), overall survival was 93% (90% CI, 85% to 100%), progression-free survival was 64% (90% CI, 46% to 88%), GVHD-free relapse-free survival was 41% (90% CI, 23% to 73%), and chronic GVHD-free relapse-free survival was 64% (90% CI, 46% to 88%). Our data indicate that mMUD PBSCT using PTCy-based GVHD prophylaxis is safe and feasible. All patients engrafted, and rates of NRM (7%) and acute GVHD (15%) at 1 year were low. There was only 1 case (3%) of severe CRS. Compared with previously published outcomes, mMUD PBSCT using PTCy-based GVHD prophylaxis has a safety and efficacy profile that may not be different from that of PBSCT from matched donors. These results further solidify that all patients who require blood or BM transplantation should be able to find an acceptable donor.