Precision medicine has significant potential, but the complexity of cancer has limited its success
Advances in genomic sequencing and the massive expansion in treatment options over the past two decades are changing our approach to cancer treatment. Whereas older therapeutic strategies relied on non-cancer-specific mechanisms of action, molecular profiling has supported the rise of targeted therapies. By examining expression patterns within a cancer relative to healthy tissues, researchers have been able to develop drugs that are more precise, resulting in fewer side effects for the patient.
While precision medicine has revolutionized cancer therapy, it remains far from the “silver bullet” many hoped it would be. The heterogeneous nature of cancer coupled with the complex interplay between the tumor and its microenvironment has attenuated the potential of this approach. Many patients fail to respond to therapy, while others develop resistance during treatment.
With only 7% of cancer patients benefiting from genome-informed therapeutic approaches, researchers are now exploring alternative avenues to realize the full potential of precision medicine.
Model systems enable functional drug testing at the patient level
To develop more effective precision medicine strategies, a multitude of drug candidates must be evaluated against patient-specific cancers. From a logistical, economic, and ethical standpoint, this is impossible to do in humans at the clinical trial level. Instead, researchers have turned to functional drug testing in model systems of cancer.
These model systems need to consider the composition, organization, and function of the patient-derived cancer tissue and its microenvironment to faithfully recapitulate clinical response. Unsurprisingly, creating these model systems for functional drug testing is highly complex, resulting in the adoption of a number of approaches to recapitulate patient biology ex vivo.
Direct profiling involves evaluating drug efficacy against freshly biopsied tumor samples. Because the material is minimally processed and accessed quickly post-biopsy, native features within the tumor microenvironment, such as immune cell populations and tumor expression, are more faithfully preserved compared with other methods requiring greater ex vivo processing. Despite these desirable characteristics, direct profiling studies can be limited in scope due to the small amounts of available biopsy material.
Patient-derived xenografts (PDXs) offer a partial solution to the shortcomings of direct profiling. PDX models are developed by implanting patient-derived tumor tissue in immune-deficient mice. The tumor then subsequently grows within the mouse, enabling researchers to expand the patient’s minimal tumor material and perform in vivo functional drug testing. However, variable implantation success rates and the longer timeframes required to allow for in vivo tumor growth can mean that this method can sometimes be suboptimal for developing individualized therapeutic strategies, for example, in patients with aggressive disease who require more immediate treatment options.
Patient-derived organoids (PDOs) occupy a middle ground between the above-mentioned methods. These models self-organize into 3D structures to better recapitulate the tumor’s structure and function when compared with traditional 2D cell culture. While they cannot fully recapitulate native immune cell population seen in direct profiling methods, they have the distinct advantage of allowing researchers to greatly expand patient-derived tumor cell populations and perform high-throughput screening assays. These assays can yield valuable information regarding tumor vulnerabilities, mechanisms of resistance, and potential therapeutic strategies. While each of these model systems has its unique limitations, one constant remains — they have all demonstrated translational success. To learn more about these models and their roles in functional drug testing and clinical outcomes, see our article Improving precision stratification using patient-derived model systems.
