Future Trends in Brain Tumor Diagnosis

0
154

The clinical practice of brain tumor diagnosis is entering a period of dramatic change and stands to be profoundly transformed over the next decade by advances in our understanding of tumor biology and emerging technologies. We have already begun to appreciate the effects of an increasingly sophisticated understanding of genetics in the way our tumor patients are treated. The implementation of emerging tools, such as artificial intelligence and liquid biopsy, hold promise for making advances in genetics and epigenetics a routine part of clinical practice that is widely available to the entire population of brain tumor patients.

Genes/Tumors are Not All Equal

The cancer genome atlas project ushered in a new era in the diagnosis of gliomas when a comprehensive survey of the genetic makeup of human glioblastomas was published in 2008. Since that time, several landmark studies (Glioma Groups, Diffuse Low Grade Gliomas) have transformed the field of brain tumor diagnosis from one based primarily on microscopic tissue architecture, to one heavily reliant on molecular diagnosis. Recent work has highlighted the importance of epigenetics in the further characterization of gliomas and offered biological insight into why some patients respond to treatment better than others.

o

Sequencing is increasingly being used on a routine basis to predict prognosis and raise the possibility for therapies targeted at specific genetic abnormalities. The NCI-MATCH trial is unique in that it enrolls patients with tumors of diverse origin (i.e., colon, rectal, breast, lung, brain, prostate) and assigns them to therapy based on their common targetable genetic abnormalities. It is likely that clinical trials in the future will follow the lead of the NCI-MATCH trial and enroll patients based on molecular status to eliminate the noise of molecular heterogeneity amongst study subjects.

The Need for New Approaches

A gap exists between our broadening understanding of the genetics of brain tumors and the options currently available for treatment. Development of targeted agents tailored for central nervous system activity will be essential in leveraging our progressively sophisticated understanding of the molecular underpinnings of glioma into advances in both diagnosis and precision medicine treatment for brain tumor patients.

A related challenge in delivering a precise, medicine-based diagnosis and treatment of brain tumors on a population and global scale relates to the cost, logistics and technical expertise inherent to molecular diagnosis. While facilities for routine histology exist in almost all centers where brain tumor surgery is performed, access to state-of-the-art molecular diagnosis tends to be confined to tertiary and quaternary centers. The solution to this disparity may lie in the implementation of artificial intelligence. For example, based on routine histology alone, researchers at New York University have demonstrated that genetic driver mutations in lung cancer can be predicted using Google’s inception v3 deep convolutional neural network. In this manner, artificial intelligence may become an essential tool that links advances in molecular diagnosis to conventional histology.

The Future in Optics

For the foreseeable future, microscopic tissue diagnosis will still play a central role in the diagnosis of brain tumor surgery. That said, a class of emerging optical technologies may disconnect the pathway for tissue diagnosis from the pathology lab. Light sheet microscopy, microscopy with ultraviolet surface excitation (MUSE) and stimulated Raman histology offer three examples of how advances in optics can be leveraged to create diagnostic grade histologic images without the need for tissue sectioning in a pathology laboratory.1 Each of these technologies has the potential for generating natively digital histologic images, which can be quantitatively analyzed via artificial intelligence algorithms. Slide-free histology is inherently less resource intensive than a conventional histology lab and may ultimately be deployed in centers that lack the resources and/or expertise to run, maintain or support a conventional histology lab. Some slide-free methods, like stimulated Raman histology, can also be deployed in the operating room, putting histologic data at the fingertips of the surgeon and informing decision-making. With emerging optical technologies, barriers between histology and the OR are likely to fall.

The Liquid Biopsy

Tissue-free diagnostics, such as liquid biopsy, also hold the potential to revolutionize the way we screen, diagnose, monitor and treat brain tumors. The shortcomings of radiographic diagnosis, including the overlapping appearance of disparate lesions and the inability to differentiate tumor progression and treatment effect, have set the stage of improvements in diagnostic tools. Moreover, the risks associated with obtaining tissue from the brain via traditional brain biopsy preclude serial testing that might be ideal in following treatment response in the face of equivocal radiographic data. Emerging targets for liquid biopsy with relevance to brain tumor diagnosis include:

  • Circulating tumor cells
  • Cell-free fractions
  • Extracellular vesicles

Each of these targets hold potential to provide a window into evolving brain tumor genetics over the course of treatment that would historically be possible only with tissue biopsy. These targets also face unique logistic and regulatory hurdles along the path to mainstream clinical use. Nonetheless, given the clear need for less invasive, more accurate ways for monitoring brain tumors, adoption will grow in years to come.

2030 and Beyond

Imagine a time when:

  • The diagnosis of a specific genetic subtype of a brain tumor is made without violating the skull.
  • Treatment interventions are targeted to a specific subtype, less chemotherapy then directed destruction.
  • Survival after diagnosis with a malignant brain tumor begins to mirror the successes already seen with other solid tumors.

That is a future we can all hope for.

References
1. Orringer, D. A., Pandian, B., Niknafs, Y. S., Hollon, T. C., Boyle, J., Lewis, S., . . . Camelo-Piragua, S. (2017). Rapid intraoperative histology of unprocessed surgical specimens via fibre-laser-based stimulated Raman scattering microscopy. Nature Biomedical Engineering, 1(2), 0027.

[aans_authors]

Print Friendly, PDF & Email
o