2012 Young Investigator Award – Final Report

Background on my research:

The broad objective of my research is to analyze data from completed clinical trials in cancer patients in order to learn something about the disease and/or drug effects that can inform the future development and/or use of cancer therapeutics. Specifically, I do this by developing disease progression models, which are mathematical models that describe disease severity over time and allow one to estimate the effect of a therapy on disease severity. Any quantitative measure of disease severity that is recorded at various time points during therapy can be modeled in this manner. Examples in oncology include tumor size on cross-sectional imaging, tumor markers in the blood such as prostate specific antigen, and symptom scores that are reported directly by patients.

Background on this project:

I worked with a data set from CALGB 40502, a phase III trial that randomized patients with metastatic breast cancer to one of three treatments: paclitaxel, nab-paclitaxel, and ixabepilone. The primary results of the trial have already been reported, and paclitaxel was found to be superior to the other treatments with respect to both efficacy and safety. All three of the drugs work by a similar mechanism of action and are known to cause neurotoxicity, also known as chemotherapy-induced peripheral neuropathy (CIPN). During the trial, patients filled out a validated questionnaire about symptoms related to neurotoxicity every 28 days, which corresponded with the first day of each cycle of chemotherapy. I used the longitudinal data from these questionnaires, along with dosing data and baseline clinical and demographic data, to develop model of CIPN that describes the trajectory of this “disease” over time. Using this model with early data from patients on therapy, we are able to predict (with some precision) which patients are more likely to develop severe CIPN with additional therapy.

Progress on this project:

The study is now complete and results were presented at the American Society of Clinical Oncology (ASCO) Annual Meeting in June 2015 in Chicago (please see attached poster). I am now preparing the manuscript and hope to submit it for publication before the end of 2015.

ASCO Presentation
Other didactic training/career development activities:

I completed a Master’s in Pharmacometrics at the University of Maryland in 2014. I have also been actively involved in clinical research as an investigator for both investigator-initiated and industry-sponsored studies. Finally, I have been awarded a Translational Scholar Award in Pharmacogenomics and Personalized Medicine (K23) from the National Institute of General Medical Sciences (NIGMS) for my project entitled “Genomic prediction tools developed using phenotypes from disease progression models”. This is a four year award, with Drs. Nancy Cox and Mark Ratain as my co-mentors, in which I have proposed to undertake additional training in genomics and statistical genetics. My K23 Award period started on May 1, 2015.

Publications (during the award period):

Maitland ML*, Wu K, Sharma MR*, Jin Y, Kang SP, Stadler WM, Karrison TG, Ratain MJ, Bies RR. Estimation of renal cell carcinoma treatment effects from disease progression modeling. Clin Pharmacol Ther. 2013 Apr;93(4):345-51. *co-primary authors

Sharma MR, Karrison TG, Kell B, Wu K, Turcich M, Geary D, Kang SP, Takebe N, Graham RA, Maitland ML, Schilsky RL, Ratain MJ, Cohen, EEW. Evaluation of food effect on pharmacokinetics of vismodegib in advanced solid tumor patients. Clin Cancer Res. 2013 Jun 1;19(11):3059-67.

Xu JM, Wang Y, Ge FJ, Lin L, Liu ZY, Sharma MR. Severe irinotecan-induced toxicity in a patient with UGT1A1*28 and UGT1A1*6 polymorphisms. World J Gastroenterol. 2013 Jun 28;19(24):3899-903.

Akthar AS, Golden DW, Nanda R, Sharma MR, Te HS, Reddy KG, Zhang X, Malik R.Early and Severe Radiation Esophagitis Associated With Concurrent Sirolimus. J Clin Oncol. 2014 Jul 14.

Xu JM, Liu XJ, Ge FJ, Lin L, Wang Y, Sharma MR, Liu ZY, Tommasi S, Paradiso A. KRAS mutations in tumor tissue and plasma by different assays predict survival of patients with metastatic colorectal cancer. J Exp Clin Cancer Res. 2014 Dec 10;33(1):104.

Sharma MR, Gray E, Goldberg RM, Sargent DJ, Karrison TG. Resampling the N9741 Trial to Compare Tumor Dynamic Versus Conventional End Points in Randomized Phase II Trials. J Clin Oncol. 2015 Jan 1;33(1):36-41.

Book Chapters:

Sharma MR, Polite BP. Fundamentals of DNA Synthesis. In W.M. Stadler (Ed.), Cancer Biology

Review: A Case Based Approach (Demos Medical Publishing, 2013).

Sharma MR, Ratain MJ. Pharmacology. In JF Holland and E Frei III (Ed.), Cancer Medicine

(Wiley Publishing, in press).

The broad objective of my research is to analyze data from completed clinical trials in cancer patients in order to learn something about the disease and/or drug effects that can inform the future development and/or use of cancer therapeutics. Specifically, I do this by developing disease progression models, which are mathematical models that describe disease severity over time and allow one to estimate the effect of a therapy on disease severity. Any quantitative measure of disease severity that is recorded at various time points during therapy can be modeled in this manner. Examples in oncology include tumor size on cross-sectional imaging, tumor markers in the blood such as prostate specific antigen, and symptom scores that are reported directly by patients.

I am working with a data set from CALGB 40502, a phase III trial that randomized patients with metastatic breast cancer to one of three treatments: paclitaxel, nab-paclitaxel, and ixabepilone. The preliminary results of the trial have already been reported, and paclitaxel was found to be superior to the other treatments with respect to both efficacy and safety. All three of the drugs work by a similar mechanism of action and are known to cause neurotoxicity, also known as chemotherapy-induced peripheral neuropathy (CIPN). During the trial, patients filled out a validated questionnaire about symptoms related to neurotoxicity every 28 days, which corresponded with the first day of each cycle of chemotherapy. I am using the longitudinal data from these questionnaires, along with dosing data and baseline clinical and demographic data, to develop models of CIPN that describes the trajectory of this “disease” over time. By understanding the factors that impact the between patient variability with respect to CIPN, we will be able to predict (with some precision) which patients are more likely to develop severe CIPN more quickly. Simulations from the model will be done to explore which candidate endpoints (e.g., neuropathy score after 3 months of therapy) would be most informative to measure in future clinical trials of drugs that cause, prevent, or treat neuropathy.

Progress:

The initial work involved cleaning the data and deciding on the phenotype of interest. The selected phenotype has been termed “fgsum4” and refers to the sum of the scores for four sensory peripheral neuropathy indicators (numbness and tingling in the hands and feet; discomfort in the hands and feet). Exploratory analyses of the data revealed significant between patient variability in the time course and severity of CIPN. CIPN was generally less severe with paclitaxel than with nab-paclitaxel or ixabepilone. I have developed population kinetic-pharmacodynamic models using longitudinal data from individuals on dosing and the fgsum4 phenotype, with a separate model for each chemotherapy drug. I am currently working to evaluate and improve these models before exploring the potential impact of a number of baseline covariates: height, weight, age, race. I presented an update of the models at the Alliance Population Pharmacology and Pharmacogenomics Committee meeting on November 8, 2013. I anticipate that the models will be finalized in 2014, before the end of the award period, at which time I will be able to present them at a major national meeting and prepare a manuscript for publication. Future goals are to improve these models by validating them with independent data sets and to use model-estimated phenotypes to make pharmacogenomic discoveries. I have discussed my ongoing work on this project and broader aspects of my training with my mentor, Dr. Mark Ratain, at least monthly. I have also organized teleconferences at least every other month with a larger group of advisors, including Dr. Joga Gobburu (University of Maryland) and Dr. Kouros Owzar (Duke).

Other didactic training/career development activities:

I have completed 11 courses toward a Master’s in Pharmacometrics at the University of Maryland (http://www.pharmacy.umaryland.edu/academics/pharmacometrics/courses.html) over the past 16 months (since Fall 2012). I am currently enrolled in the final course and will complete the degree in Spring 2014. My instructor for the Master’s program, Dr. Joga Gobburu, has become an advisor for my CRF YIA project and will be a member of the advisory committee for my proposed K23 award. I have also been actively involved in clinical research as an investigator for both investigator-initiated and industry-sponsored studies. Finally, I have submitted an application for the Translational Scholar Award in Pharmacogenomics and Personalized Medicine (K23) entitled “Genomic prediction tools developed using phenotypes from disease progression models” with Dr. Nancy Cox and Dr. Mark Ratain as co-mentors, in which I have proposed to undertake additional training in genomics and statistical genetics. If my application is successful, I would start my K23 training on July 1, 2014.