Twenty-first century health care has a language problem. Wed to our traditions, we continue to discuss, research and treat disease using what should be a dead language, one based on gross and microscopic anatomy. In an age of molecular medicine our words obscure our communication and our understanding.

Through history, we described disease based on how it looked. We called tuberculosis “consumption” due to the weight loss with which it was associated. Hippocrates coined the term “cancer” from “carcinos,” Greek for crab, an animal whose claws spread from a central core, the way many tumors do. Later, microscopes facilitated the science of histology or microanatomy, which replaced gross anatomy as our primary tool for identifying and classifying disease.

Nineteenth and 20th century medicine was based in large part on a process of identifying an abnormal appearance and experimenting with methods to return that appearance back to normal. But advances in science, particularly in chemistry, molecular biology and genetics, suggested that appearance alone is not sufficient. An early sign came in 1960, when Penn’s Peter Nowell discovered an early cancer biomarker unique to chronic myelogenous leukemia. More recent work has identified specific molecular and genetic markers that correlate with a tumor’s actual biology, suggesting how that tumor survives and grows and therefore directing us to possible treatment.

Our delineation of molecular pathways, the submicroscopic processes that control both normal function and disease, blurs the static, picture-based logic of diagnosis and treatment of disease. We can now use drugs that block vessel formation to choke off the blood supply of a tumor rather than surgically remove or blast toxic amounts of radiation at the organ in which the tumor lives. We can design targeted “smart bombs” that carry poisonous payloads that destroy the tumor but spare the normal tissue around it.

This type of progress in science has been spectacular, and much of ithas translated to improvements in patient care. Unfortunately, some of that progress has been stalled—partly because our language is too imprecise to reflect the accuracy with which we can now understand, and ultimately treat, disease.

Take lung cancer as an example. The presence or absence of a marker called a KRAS mutation correlates with resistance to a category of drug frequently used to treat the most common form of lung cancer. A similar relationship exists between this particular KRAS mutation and colorectal cancer. Connections between other tumors and molecular and generic markers emerge frequently. Ultimately, we will define cancer biology and treatment based on a complex signature of molecular activity. Traditionally, we treated all lung cancers or all colon cancers lumped together.

For now, we still largely use traditional, imprecise anatomic and histologic categorizations. As an investor, I see the inefficiencies in business plans that evolve from this imprecision. Pharmaceutical and biotechnology companies design trials for anatomic diseases. Diagnostic companies design and market tests that use anatomic site as the primary diagnosis and molecular signature only as a means of sub-categorization. And for now the FDA has done little to encourage change.

Hopefully, as the accuracy of our definitions increases, the efficiency of drug development and disease prevention methods will improve, and incorporation of precise definitions based on disease function, rather than anatomic location or histologic appearance, will realize the full benefits of personalized medicine.