A new appreciation of the pathobiological foundation of mucositis, and the application of genomics to risk assessment, heralds an individualized and more effective approach to intervention for this costly, often disabling, toxicity, according to specialists who spoke at a session on mucosal injury during the 2012 Annual Meeting of the American Society of Clinical Oncology (ASCO).
New Mechanistic Understanding
Old concepts are being replaced by a pathobiological paradigm that better captures the underlying mechanism and allows for more effective interventions. Mucositis is now “about bioinformatics, targeted treatment, risk prediction, and other exciting frontiers,” said Douglas Peterson, DMD, PhD, of the University of Connecticut Health Center in Farmington.
Stephen T. Sonis, DMD, DMSc, of Biomodels in Watertown, Massachusetts, described the biological cascade leading to mucosal injury as one involving the mesenchyme, extracellular matrix, and epithelium; the local environment; and the microflora. Once thought of as a direct epithelial process in which chemotherapy or radiotherapy killed cells, the development of mucositis is actually a multifactorial process within the diverse tissues of the submucosa, he said.
“The cascade of events starts with oxidative stress, followed by an innate immune response, leading to a primary damage response, initiation of signaling pathways, amplification and feedback, then ulceration, surface colonization, and healing,” Sonis said.
The involvement of so many processes means, he said, “that in developing drug targets and in looking for risk predictors, we have a bigger palette to go after, versus just focusing on epithelial cells, as in the historical paradigm.”
Emerging areas of research are the local environment of the gastrointestinal tract and mouth, especially salivary and intestinal proteins secreted in response to mucosal injury, as well as changes in bacterial flora, Sonis said.
Targeted Agents May Not Be Much Better
Dorothy Keefe, MD, MBBS, of Royal Adelaide Hospital in Australia, cautioned against thinking that new targeted agents will carry less risk of mucositis. The mTOR inhibitors, for example, produce a unique mucositis pattern.
“Every new drug class brings new toxicities, despite the temptation to think that ‘targeted’ equals ‘nontoxic,’ ” she said. The toxicities of targeted therapies are often clinically significant, long-lasting, and not reversible, she added.
Keefe pointed out that signaling pathways may be related to both the toxicity and oncogenesis, and interfering with pathways is “tricky territory.” Understanding of mechanisms, she said, is key to optimal management.
“Time-tested empiric foundational approaches” that have been used for decades will continue to be important in the new frontier of targeted cancer therapies and their toxicities, Peterson said. These include patient education and compliance, simple wound care approaches, structured approaches to pain control, adequate hydration and nutritional support, surveillance, and treatment of infectious complications.
“The problem is that these approaches are not fundamentally targeting the pathobiology,” he said. “The new world order, where we are headed, will include individualized risk profiles and customized therapies.”
“Personalized cancer medicine involves both tumor and toxicity. It is not good enough to personalize the cancer treatment. We have to personalize the supportive care as well, and risk prediction for toxicity is becoming a reality,” Keefe said.
“We are at a very serendipitous time,” Sonis agreed. “Risk prediction has become possible through the evolution of genomics.”
Single-nucleotide polymorphisms (SNPs) are the most common form of variation in the genome. SNPs can reside intracellularly (and affect transcription) or extracellularly (and be an indicator of toxicity risk).
Advances in bioinformatics are making it possible to select clusters of SNPs, out of millions, that can predict for toxicity. This is already possible for patients undergoing conditioning regimens for autologous stem cell transplant, 40% of whom will develop severe mucositis and 60% of whom will therefore be treated unnecessarily, he said.
In a retrospective analysis of myeloma and lymphoma patients undergoing transplant at a single center, Sonis and his team identified 51 patients who developed oral mucositis and 102 who did not. They looked for SNPs in the DNA of patients’ saliva and found that a cluster of 82 SNPs that predicted for mucositis with 99.3% accuracy.
The area under the ROC (receiver operating characteristic) curve was 0.997 (0.9-1.0 is considered an excellent predictor). The ROC for mammography is 0.74, he noted for comparison. A multicenter study of the model is planned, and this model is being further developed for commercialization.
“We were thrilled,” Sonis commented. “This approach should be a game changer in terms of our ability to predict mucositis and other biologically based toxicities. This means we can be much more precise in developing risk assessments, understanding the patient’s response to risk, and being able to favorably impact toxicity with targeted interventions for the patient at risk.”