Two years ago, Professor Fabrice André and colleagues published a provocative article in Nature, calling for an end to the traditional practice of classifying metastatic cancers by their organ of origin. Instead, they advocated for a new framework, grouping tumours according to their molecular characteristics, an approach that, they argue, would better align with the realities of modern cancer therapeutics.
Rethinking Metastatic Cancer Classification is important for advancing cancer treatment strategies.
The rationale behind this proposal is compelling. Most new cancer drugs are designed to target specific molecules or genetic mutations, irrespective of where the cancer began. Sequential drug trials, conducted organ by organ, have led to frustrating delays in patient access to potentially life-extending therapies. Examples such as nivolumab and olaparib highlight how patients with the “wrong” cancer type, but the “right” molecular marker, have waited years for treatments that could have helped them much sooner.
André’s vision is ambitious and, at first glance, logical. If treatments act on molecular targets, why not classify and treat cancers accordingly? The argument extends beyond drug approval timelines: current hospital structures, clinical education, and patient support systems are all based on the primary site of disease, which may be increasingly outmoded in an era of precision medicine.
However, this proposal has not gone unchallenged. Critics, including leading pathologists and oncologists, point out that the organ of origin does not merely provide a convenient label. The tissue context can shape both tumour behaviour and treatment response. A drug effective against a particular mutation in one organ may be less effective, or even ineffective, in another due to differences in microenvironment, drug metabolism, and interaction with surrounding tissues.
There are also practical and social complexities. For many patients, the identity of their cancer, be it breast, lung, or bowel, provides a source of community and support. To strip away these distinctions risks undermining structures that offer psychological and practical help. Moreover, the infrastructure for comprehensive molecular testing is not universally available, raising concerns about equity and access in lower-resource settings.
Some critics suggest that, rather than a revolution, what is needed is evolution: a hybrid model that integrates molecular profiling with the clinical knowledge and expertise built up around specific cancer types. Indeed, Professor André himself has since acknowledged that tumour origin remains relevant in certain scenarios, and that a scoring system combining both approaches may be the way forward.
The future may well lie in a more nuanced approach, embracing both the promise of precision medicine and the hard-won wisdom of traditional oncology.
Manuela Boyle March 2026
