Complete hazard ranking to analyze right-censored data: An ALS survival study

Huang, Zhengnan; Zhang, Hongjiu; Boss, Jonathan; Goutman, Stephen A.; Mukherjee, Bhramar; Dinov, Ivo D.; Guan, Yuanfang

doi:10.1371/journal.pcbi.1005887

Cited by 31 publications

(47 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This defined the outcome variable more precisely, which led to a more adequate training of regression models (here: Gaussian process models) explaining the algorithm's superior prediction performance. In line with these results, the strategy outperformed a standard Cox model by 20% accuracy 22 .…”

Section: Comparative Assessment Of Prediction Methodssupporting

confidence: 74%

“…This analysis revealed several features, already well reported in the literature, such as age, gender [22][23][24] and respiratory capacity. 22,29,44,45,46 which were strong predictors of survival, but were less informative for predicting disease progression, whereas other features, including limb motor function ALSFRS-R scores (hands and legs function) and specific ALS staging scores 47 were more informative for predicting disease progression rather than survival. Creatinine, already suggested before as a measure predictive of ALS prognosis 16,29,48,49 , 50 was also found to be predictive in this challenge, and interestingly this was specific for patients early in their disease.…”

Section: Discussionsupporting

confidence: 66%

“…The most frequently used features for all sub-challenges were those that are well described in the literature as being strongly related to ALS prognosis: time from disease onset and total ALSFRS score 23,24,25 . Age and gender were also frequently used but were more informative for predicting survival rather than disease progression, in line with literature 22,23,24,26,27,28 . While age and gender were generally predictive in this data as well, they were not specifically more predictive for one particular subgroup of patients, and therefore less relevant for stratification, while still not deterring from its overall importance as a predictor.…”

Section: Predictive Clinical Featuressupporting

confidence: 65%

“…First, only few (3%) of these "slow progressing" patients had bulbar onset. While bulbar onset had been frequently correlated with poorer prognosis [22][23][24] , this cluster analysis suggests that bulbar patients will rarely be classified as a slow progressing disease presentation. A second important observation is low creatinine levels (average of 67 mcmol/L in the first three month of data collection, compared to a normal desired range of 74.3-107mcmol/L).…”

Section: Clinically Relevant Patient Clustersmentioning

confidence: 83%

See 3 more Smart Citations

Stratification of amyotrophic lateral sclerosis patients: a crowdsourcing approach

Kueffner

Zach

Bronfeld

et al. 2018

Preprint

View full text Add to dashboard Cite

Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disease with substantial heterogeneity in clinical presentation with an urgent need for better stratification tools for clinical development and care. In this study we used a crowdsourcing approach to address the problem of ALS patient stratification. The DREAM Prize4Life ALS Stratification Challenge was a crowdsourcing initiative using data from >10,000 patients from completed ALS clinical trials and 1479 patients from community-based patient registers. Challenge participants used machine learning and clustering techniques to predict ALS progression and survival. By developing new approaches, the best performing teams were able to predict disease outcomes better than currently available methods. At the same time, the integration of clustering components across methods led to the emergence of distinct consensus clusters, separating patients into four consistent groups, each with its unique predictors for classification. This analysis reveals for the first time the potential of a crowdsourcing approach to uncover covert patient sub-populations, and to accelerate disease understanding and therapeutic development.Amyotrophic lateral sclerosis (ALS) is a neurodegenerative disorder which causes the death of motor neurons that control voluntary muscles. The loss of motor neurons leads to progressive muscle weakening and paralysis and on average patients will survive only 3-5 years from symptom onset 1 . Despite being known for over 150 years, we only have limited understanding of the biological mechanisms underlying ALS and existing therapeutic options merely extend survival by a few months 2,3 . One of the biggest challenges in ALS treatment and research today is the wellestablished heterogeneity of the disease 4,1 ; ALS patients can have widely different patterns of disease manifestation and progression, and genetic analyses suggest heterogeneity of the underlying biological mechanisms as well 5,6,7,8 . This heterogeneity has detrimental effects on clinical trial planning and interpretation, as it might mask drug effects 3 , on attempts to uncover disease mechanisms, and on clinical care, as it increases uncertainty about prognosis and makes treatment course planning challenging. Thus, successfully stratifying ALS patients into clinically meaningful sub-groups can be of great value for advancing the development of effective treatments and achieving better care for ALS patients.Early classification systems for ALS patients were based on clinical presentation of the disease and were intended for ascertainment of an ALS diagnosis, but were limited in their ability to predict disease prognosis or suggest underlying disease mechanisms 9,10,4 . More recent attempts of ALS patient classification focused on prediction of clinical outcomes but were often limited by small sample sizes lacking the highly needed detailed characterization of patient subgroups 11,12,13,14 . In the current study, we sought to use the power of state of the art machine learning algorith...

show abstract

Section: Comparative Assessment Of Prediction Methodssupporting

confidence: 74%

Section: Discussionsupporting

confidence: 66%

Section: Predictive Clinical Featuressupporting

confidence: 65%

Section: Clinically Relevant Patient Clustersmentioning

confidence: 83%

See 2 more Smart Citations

Stratification of amyotrophic lateral sclerosis patients: a crowdsourcing approach

Kueffner

Zach

Bronfeld

et al. 2018

Preprint

View full text Add to dashboard Cite

show abstract

“…Figure 3A-B shows that the network activity preserves the large-scale patterns present in the gene expression data and appears much less noisy. For each high-risk clinical subtype, we classified the 30% of patients that were highest risk by GuanRank 30 as truly high risk and classified the other 70% as low-risk. We then randomly split the samples into a training and test set ( i .…”

Section: Resultsmentioning

confidence: 99%

Genetic program activity delineates risk, relapse, and therapy responsiveness in Multiple Myeloma

Wall

Turkarslan

et al. 2020

Preprint

View full text Add to dashboard Cite

38Despite recent advancements in the treatment of multiple myeloma (MM), nearly all patients 39 ultimately relapse and many become refractory to their previous therapies. Although many 40 therapies exist with diverse mechanisms of action, it is not yet clear how the differences in MM 41 biology across patients impacts the likelihood of success for existing therapies and those in the 42 pipeline. Therefore, we not only need the ability to predict which patients are at high risk for 43 disease progression, but also a means to understand the mechanisms underlying their risk. We 44 hypothesized that knowledge of the biological networks that give rise to MM, specifically the 45 transcriptional regulatory network (TRN) and the mechanisms by which mutations impact gene 46 regulation, would enable improved predictions of disease progression and actionable insights for 47 treatment. Here we present a method to infer TRNs from multi-omics data and apply it to the 48 generation of a MM TRN that links chromosomal abnormalities and somatic mutations to 49 downstream effects on gene expression via perturbation of transcriptional regulators. We find that 50 141 genetic programs underlie the disease and that the activity profile of these programs fall into 51 one of 25 distinct transcriptional states. These transcriptional signatures prove to be more 52 predictive of outcomes than do mutations and reveal plausible mechanisms for relapse, including 53 the establishment of an immuno-suppressive microenvironment. Moreover, we observe subtype-54 specific vulnerabilities to interventions with existing drugs and motivate the development of new 55 targeted therapies that appear especially promising for relapsed refractory MM. 56 57 Introduction 58 59 Multiple Myeloma (MM) is a cancer of malignant plasma cells in the bone marrow (BM) that has 60 a prevalence of approximately 86,000 new cases per year. 1 Several clinical subtypes of MM have 61 been established on the basis of characteristic cytogenetic features, including various 62 translocations, gain or loss of chromosomal arms, deletion of specific chromosomes, and 63 hyperdiploidy. 2-4 Accordingly, MM is a complex disease of great heterogeneity that exhibits64 subtype-specific drivers of progression. 5,6 Efforts to better characterize the biology and 65 therapeutic vulnerabilities of MM have increased exponentially in recent years, as can be seen 66 from the number of research articles, clinical trials, and public availability of matched genomic, 67 transcriptomic, and patient data. However, the myriad combinations of chromosomal aberrations 68 and somatic mutations, coupled with the complex dependence of MM progression on the BM 69 microenvironment, have precluded a mechanistic understanding of the disease on a patient-70 specific level. 72The 5-year survival rate of MM is approximately 50% 7 , which is nearly double what it was in the 73 late 1980s. 8 The reason for the improved outcomes is a series of therapeutic advances that 74 include the introduction of autologous stem cell t...

show abstract

Machine Learning to Predict Anti–Tumor Necrosis Factor Drug Responses of Rheumatoid Arthritis Patients by Integrating Clinical and Genetic Markers

Guan

Zhang

Quang

et al. 2019

Arthritis & Rheumatology

Self Cite

102

View full text Add to dashboard Cite

Objective. Accurate prediction of treatment responses in rheumatoid arthritis (RA) patients can provide valuable information on effective drug selection. Anti-tumor necrosis factor (anti-TNF) drugs are an important second-line treatment after methotrexate, the classic first-line treatment for RA. However, patient heterogeneity hinders identification of predictive biomarkers and accurate modeling of anti-TNF drug responses. This study was undertaken to investigate the usefulness of machine learning to assist in developing predictive models for treatment response.Methods. Using data on patient demographics, baseline disease assessment, treatment, and single-nucleotide polymorphism (SNP) array from the Dialogue on Reverse Engineering Assessment and Methods (DREAM): Rheumatoid Arthritis Responder Challenge, we created a Gaussian process regression model to predict changes in the Disease Activity Score in 28 joints (DAS28) for the patients and to classify them into either the responder or the nonresponder group. This model was developed and cross-validated using data from 1,892 RA patients. It was evaluated using an independent data set from 680 patients. We examined the effectiveness of the similarity modeling and the contribution of individual features.Results. In the cross-validation tests, our method predicted changes in DAS28 (ΔDAS28), with a correlation coefficient of 0.405. It correctly classified responses from 78% of patients. In the independent test, this method achieved a Pearson's correlation coefficient of 0.393 in predicting ΔDAS28. Gaussian process regression effectively remapped the feature space and identified subpopulations that do not respond well to anti-TNF treatments. Genetic SNP biomarkers showed small contributions in the prediction when added to the clinical models. This was the bestperforming model in the DREAM Challenge. Conclusion. The model described here shows promise in guiding treatment decisions in clinical practice, based primarily on clinical profiles with additional genetic information.

show abstract

Complete hazard ranking to analyze right-censored data: An ALS survival study

Cited by 31 publications

References 48 publications

Stratification of amyotrophic lateral sclerosis patients: a crowdsourcing approach

Stratification of amyotrophic lateral sclerosis patients: a crowdsourcing approach

Genetic program activity delineates risk, relapse, and therapy responsiveness in Multiple Myeloma

Machine Learning to Predict Anti–Tumor Necrosis Factor Drug Responses of Rheumatoid Arthritis Patients by Integrating Clinical and Genetic Markers

Contact Info

Product

Resources

About