Crowdsourcing Molecular Knowledge: A Practical and Affordable Approach to Create a Molecular Tumor Board

The advent of precision medicine has revolutionized cancer treatment by tailoring therapies to individual patients based on the molecular characteristics of their tumors. The molecular somatic sequencing of tumors and germinative DNA through next-generation sequencing (NGS) has become a standard practice in cancer care.1 The interpretation of the results of these tests is not always straightforward, requiring the expertise of several professionals, such as molecular biologists, pathologists, clinical oncologists, and geneticists. In most reference cancer centers, these professionals meet periodically at molecular tumor boards (MTBs), where clinical cases with complex genomic information are comprehensively discussed.2 In developing countries such as Brazil, where most health institutions care for underserved patients in the Brazilian Unified Health System (Sistema Único de Saúde, SUS, in Portuguese), financial constraints limit simultaneous access to these professionals in an MTB. Therefore, the cases of some patients with complex genomic information do not undergo a multidisciplinary discussion due to the lack of an institutional MTB.

We initiated a crowdsourcing approach to molecular knowledge by assembling a group of Brazilian physicians from multiple cities and institutions into a virtual WhatsApp (Meta Platforms, Inc.) group. The group consisted of medical oncologists (n = 68), radiation therapists (n = 4), oncogeneticists/molecular biologists (n = 3), radiologists (n = 3), nuclear medicine personnel (n = 3), oncologic surgeons (n = 5), gynecologic oncologists (n = 3) and pathologists (n = 2). Most of the cases discussed were of breast cancer (6), followed by lung (5) and colorectal cancers (4) (Table 1).

Patient confidentiality was safeguarded by prohibiting the sharing of identifiable information. Treating physicians were instructed to summarize cases using de-identified data only, and no patient names, initials, dates of birth, or hospital registration numbers were allowed. Consent for molecular testing and subsequent discussion was obtained by the treating physicians as per institutional standards.

Discussions followed a semistructured format. Treating physicians presented molecular and clinical data in writing, and other members contributed asynchronously with their opinions, literature references, or management suggestions. While the exchanges were not moderated in real time, the treating physician summarized the discussion to ensure clarity and to highlight potential management strategies.

For each case, management options were suggested based on the genomic findings and clinical context provided by the treating physician. However, apart from some cases in which the treating physician shared treatment decisions and outcomes voluntarily, we did not prospectively follow whether these recommendations were implemented in the clinical practice, nor did we collect systematic information on patient outcomes.

From September 2022 to May 2024, we discussed 22 cases, with an average of 3.3 patients discussed per month (Table 1). The group suggested management options for all cases to the treating physicians.

As an illustration, we discussed the case of a 65-year-old female patient previously diagnosed with metastatic lung cancer carrying a BRCA2 mutation that had spread to the central nervous system and was treated with olaparib after failing cisplatin-based chemotherapy and brain radiation therapy. Olaparib had a stabilizing effect on her tumor for approximately 30 months. During a routine follow-up, a new triple-negative breast tumor was identified. Genetic testing revealed the presence of two separate BRCA2 gene mutations in the breast tumor. The initial mutation (p.Val220Ilefs4) caused a frameshift, producing a truncated and non-functional BRCA2 protein. However, the second mutation, K437fs22, corrected the reading frame of exon 11, enabling Rad51 to bind to BRCA2 properly. This new mutation restored the functionality of BRCA2, explaining the clinical resistance observed in the breast tumor. The patient underwent breast tumorectomy and radiation therapy, and she continues olaparib, with a stable disease.3

The present study showed that creating a multidisciplinary virtual group to discuss oncological cases with complex genomic information is a feasible and affordable alternative even for developing countries where financial resources are scarce to fund single institutional MTBs. The group's role was limited to generating and sharing potential management strategies; clinical implementation remained at the discretion of the treating physician. This distinction is important because, while the discussions demonstrated that a WhatsApp-based MTB is feasible, the absence of follow-up prevents us from drawing conclusions about its real-world impact on treatment decisions or patient survival.

An important strength of our model is that patient confidentiality was preserved by using only de-identified case summaries, while allowing broad access to expert input. The discussions were semistructured and conducted asynchronously, which lowered barriers to participation compared to traditional MTB meetings.

The advantages of the WhatsApp group are the low costs involved, the non-synchronic means of communication, in which the professionals may answer at any time they prefer and not need to meet at a particular time, unlike the usual MTBs, which meet either in person or virtually.

The current study has several limitations, such as its non-prospective design, the small number of cases discussed, and the lack of systematic follow-up on all patients. While some outcomes were shared by the treating physicians, a structured collection of follow-up data would strengthen future studies. Another limitation is that case presentations were not structured based solely on the abbreviated information posted by the treating physicians, and we could not ascertain if all relevant specialists saw each case, as only a few physicians in the group posted their comments on each case. Furthermore, other non-medical professionals such as pharmacists could complement the physicians' inputs and further contribute to the case discussions.

In conclusion, implementing virtual MTBs through crowdsourcing of molecular knowledge through a virtual group of physicians has proven to be a feasible and effective solution for developing countries facing financial constraints to fund single institutional MTBs. This approach has enabled comprehensive discussions of complex genomic information and has provided valuable management options for treating physicians. The case herein reported illustrates the potential of this approach, as it resulted in a successful treatment outcome for a patient with metastatic lung cancer and a secondary breast tumor carrying multiple BRCA2 gene mutations. Further research and implementation of this approach in other countries with limited access to traditional MTBs could greatly benefit cancer patients and improve their outcomes.