The Mark Foundation for Cancer Research

Patients have difficult unmet needs when standard chemotherapy produces a median survival of less than 1 year or many patients will experience severe toxicities. Blood tests can predict their survival. Analyses evaluate predictive blood tests to identify patients who often survive 1 and 2 years. A four-test model includes: albumin, absolute neutrophil count, neutrophil-lymphocyte ratio, and lymphocyte-monocyte ratio. Individual tests include: alkaline phosphatase, lymphocytes, white blood count, platelet count, and hemoglobin. Eligible patients have advanced: resistant 3rd line colorectal, and both resistant and new pancreatic and intrahepatic bile duct cancers. Eligibility characteristics include: biopsy-proven, measurable metastatic disease, NCI grade 0-2 blood tests, Karnofsky Score 100-50, and any adult age. Drugs are given at 1/4-1/3 of their standard dosages biweekly: gemcitabine, irinotecan, fluorouracil, leucovorin, and day 2 oxaliplatin every 2 weeks. In case of progression, Docetaxel is added (except colon cancer), with or without Mitomycin C, and next cetuximab (except pancreatic and KRAS BRAF mutation cancers). Bevacizumab is substituted for cetuximab in case of another progression or ineligibility. Consent was written and conforms with Helsinki, IRB, and FDA criteria (FDA #119005). Median survival is 14.5 months. Of 205 patients, 60% survive 12, and 37% survive 24 months (95% CI ± 8%). Survival is > 24, 13, and 3.8 months for patients with 0, 1-2, and 3-4 unfavorable tests, respectively. Individual "favorable and unfavorable" tests predict long and short survival. Neither age nor prior therapy discernibly affects survival. Net rates of clinically significant toxicities are less than 5%. Treatments reproduce predictable, greater than 12 and 24-month chances of survival for the aged and for patients with drug-resistant tumors. Evaluation of blood tests may change practice, expand eligibility, and personalize treatments. Findings support investigation of drug combinations and novel dosages to reverse resistance and improve safety.

Bevacizumab (bev), when added to a moderate dose combination of previously failed cytotoxins, as a third- and fourth-line therapy for refractory gastric, cholangiocarcinoma, and ovarian cancers, produced high-quality responses. The regimen was based on preclinical models designed in order to simultaneously partner both bev and each of the cytotoxins with 4-5 synergistic drugs. Eligible patients (n=9) had high-grade endometrial tumors and had failed standard chemotherapy. Bev (10 mg/kg every 2 weeks) and cyclophosphamide (150-250 mg/m2), were added to a combination of gemcitabine, fluorouracil, leucovorin, irinotecan and a platinum analogue -first without and then with docetaxel- each at approximately 1/2 to 1/3 of their standard dosage. Dose modification aimed at a repeated absolute neutrophil count (ANC) of 750-1,500 μl or platelets of 125,000-75,000 μl. Safety measures included stop-go use (intermittent, as needed, brief withholding of bev with resumption when again tolerated), of bev, and both prospective and ongoing dose modification in order to protect the bowels. Induction treatment was free of life-threatening complications. Nine consecutive patients, 3 under second- and 6 under multi-line treatment, had 9 objective responses and 8 produced long clinical benefits, 2 of which were complete responses. Seven responses created opportunities for personalized added treatment and research. Absolute median survival was 21.5 months for the 8 patients with platinum-resistant tumors. One patient was unable to tolerate a first standard adjuvant dose of paclitaxel. After rapid peritoneal progression of disease, treatment has produced 52+ months of unmaintained complete remission. Bev, in the combination that was used in this study, meets response, survival, and toxicity criteria for further testing against second- or multi-line chemotherapy-resistant tumors and also when a standard treatment is not safe.

Health information technology (IT) has great potential to transform health care and inform population health goals in clinical research, quality measurement, and public safety. To fully realize the benefits of health IT for population health, we must focus on new models that maximize efficiency, encourage rapid learning, and protect patients' privacy. In this paper we explore the advantages of a networked model for analyzing population health information, providing several examples. Although broadening the use of networked models is challenging, the societal benefits of a networked model merit continued exploration and the development of workable solutions.