Wu K, Rodrigues L, Post G, Harvey G, Miller A, Lambert L, Lopes C, Lewis B, Zou J. Concordance between dogs and humans: The use of AI in evaluating clinical cancer genomic datasets. In: American Association of Cancer Research - AACR Annual Meeting: April 8-13, 2022; New Orleans, US.
In a nutshell
- During the Annual Meeting of the American Association of Cancer Research (AACR) held in New Orleans in April 2022, researchers from Stanford and One Health Company used machine learning to analyze a clinico-genomic data set from FidoCure® Precision Medicine Platform and identify concordance between human and canine in the prognosis and predictive effect of genomic alterations.
- This is important to deepen the field of comparative oncology and affirm spontaneous cancer in dogs as an ideal ‘model’ for precision oncology, building upon recent work in tumor biology concordance.
- The machine learning analysis prognosis and predictive effect of genomic alterations in 1,303 client-owned tumor-bearing dogs identifying TP53 and PIK3CA as prognostic markers for poor survival.
- Among tumor types, hemangiosarcoma, histiocytic sarcoma and lymphoma showed the worst survival outcomes. FidoCure continues to enroll dogs for precision medicine treatment and constantly collects outcome data, therefore increasing scientific knowledge on canine precision medicine.
Although the true epidemiology of cancer in dogs is still unknown, there is an estimate that one in three dogs will develop the disease during their life. Surgery and radiation therapy are often used for local tumor control, but some patients require systemic treatment to restrain disseminated cancer cells. Treatment protocols using the maximally tolerated dose of chemotherapy are often used to control tumor metastasis and also to treat systemic tumors that can vary depending on histology type, tumor location, clinical stage and pets' health conditions.
Tumor response to standard treatment in dogs with cancer has recently become more predictable, especially for highly prevalent cancers. However, the disease-free interval and survival time of patients with high metastatic rates of tumors can be very short.
Over the last few years the treatment for humans with cancer has been expanding due to the implementation of immunotherapy and several small molecules that target specific oncogenes in cancer cells and at the same time minimize the damage to the healthy tissues. These new treatments were possible due to the continuous development and implementation of molecular biology tools in clinical practice that rapidly expanded the genomic knowledge, allowing the understanding of gene mutation effects in tumor development, therapeutic response, and prognosis of several tumors. These molecular alterations identified using high-throughput technologies are contributing to the development and approval of several target drugs by the Food and Drug Administration (FDA) for human treatment. Currently, precision medicine has been used to improve patient care in which an individual's genetic profiling guides clinical decisions and therapeutic strategies.
To address the current gap in knowledge, the researchers performed a large-scale analysis of tumor sequencing allowing the identification of genomic alterations that drive canine tumor growth and progression similar to humans. Identified drive mutations in canine tumors similar to human genetic alteration have prompted the use of FDA-approved small molecules for humans, to treat dogs, especially when the standard care treatment has not produced better outcomes for dogs. As compared to clinical trials previously conducted in dogs evaluating the efficacy of small molecules in less than 90 dogs per group, data-driven analysis using real world data can analyze thousands of dogs.
Researchers analyzed the largest clinico-genomic data set consisting of 1,293 dogs with spontaneous tumors. Single nucleotide variants and indels of the FidoCure® NGS panel were evaluated over 11 canine tumor types collected across the United States. From these dogs, treatment outcome data of 10 small molecule target therapies were also collected.
This is the first study using a high number of dogs with tumors, correlating the genetic landscape with clinical outcome. FidoCure® Personalized Platform data combines genomic information from dogs naturally affected by tumors, demographics, clinical characteristics and treatment outcomes.
Hemangiosarcoma, histiocytic sarcoma and lymphoma were correlated with shorter survival time in dogs. The researchers’ analysis identified TP53 mutations in canine tumors contributing to a worse prognosis in concordance to human specific tumors such as non-small cell lung cancer, metastatic breast cancer, and pancreatic carcinoma. TP53 is not only the most common mutated gene in human tumors, but also in dogs. PIK3CA was also correlated with a worse prognosis in dogs, and this mutation was identified in high frequency of hemangiosarcoma.
Interestingly, the researchers also identified that tumors carrying somatic BRAF mutations had a better prognosis when treated with lapatinib, ATM with vorinostat, and FLT3 with trametinib. BRCA1 germline and somatic mutations with deleterious or unknown prediction scores were correlated with better prognosis when treated with dasatinib.
- The statistical modeling reveals prognosis of canine tumor types and targeted gene mutations.
- Researchers identified association of outcomes with treatment-gene pairs allowing them to generate hypotheses for novel therapies.
- In the future the researchers aim to induce virtuous feedback loops in recommending and observing outcomes.