Abstract Tumor Growth Rates Remain Constant while Patients are on Therapy: Evidence that Continued Treatment does not Select for More Rapidly Growing Tumors Background: Increasingly, clinical investigators continue therapies with the hopes of improving survival. Where tumor growth is occurring, such a strategy will only succeed if the rate of tumor growth remains slow and unchanged while therapy continues. We tested this assumption using a large database of tumor measurements in patients on clinical trials. Results: We analyzed 2250 tumor growth curves from patients with advanced cancer and diagnoses of multiple myeloma, breast, renal, castration refractory prostate cancer (CRPC) or medullary thyroid carcinoma (MTC). Patients received various therapies: sunitinib, vandetanib, ixabepilone, capecitabine, bevacizumab, thalidomide, taxotere and interferon alpha. Measurements comprised serum markers (M-spike, PSA, CEA, calcitonin) and radiographic imaging. A simple computer-based model allowed extraction of statistically valid results for growth and decay rates (in over 85% of cases) and the fraction of surviving tumor in cases where marked responses occurred (up to 30% of MTC and CRPC). Analysis of the data obtained while patients were on therapy (including those still alive when the study closed and also those who died but survived long after therapy was discontinued) showed little evidence of an increase (0.3%) or a decrease (0.7%, most being MTC) in growth rate while they were treated, consistent with expansion of a pre-existing clone that is relatively resistant. Also in those MTC and CRPC cases where there was no detectable growth for a period of time, restoration of growth was consistent with expansion of a pre-existing, inherently-resistant clone, growing at a constant rate, rather than with emergence of a clone that has acquired resistance. In cases where therapy continued well beyond a year (median, 25%-75%: 498, 420-727 days) until just before death (interval duration median, 25% to 75%: 95, 63-129 days), there was neither evidence of acceleration nor the slowing that would be predicted by a Gompertz model. Many patients in whom therapy reduced the tumor growth rate to well below that of placebo, died sooner than expected. Modeling their survival data and their tumor growth rates suggested a return to the pre-therapy rate when treatment was discontinued. Conclusion: Data obtained on tumor growth rates for 2250 patients receiving therapy in clinical trials shows little evidence that tumor growth rates change while patients receive treatment. Analysis of the smaller sample of cases where death occurred within months after therapy ceased suggests that while growth rates remain constant while patients continue therapy, tumor growth returns to the pre-treatment rate when therapy is discontinued. These observations support a strategy of continuing therapy beyond the conventional criteria for discontinuation. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 765. doi:1538-7445.AM2012-765