In August 2014, a 57-year-old woman was diagnosed with invasive ductal carcinoma of the breast (Grade II). The tumor was hormone receptor-positive (ER+, PR+), HER2-negative, with a moderate Ki67 proliferation index (10%). Alongside her cancer diagnosis, the patient also had a comorbidity (hypertension).
Initial Treatment and Response
The patient began her treatment with a neo-adjuvant chemotherapy regimen consisting of Fluorouracil, Adriamycin, and Cytoxan (FAC), followed by surgery and 12 weeks of Paclitaxel. This course of treatment was complemented by radiotherapy and hormonal therapies, including Anastrozole and Zoledronic acid, resulting in a period of complete response.
Disease Recurrence and Progression
In 2019, five years after her initial treatment, the patient experienced a recurrence, with metastasis detected in both her liver and bones. In response, she was treated with Fulvestrant and Palbociclib, followed by additional lines of treatment, including Nab-paclitaxel, Carboplatin, and Everolimus. Despite the aggressive treatment regimen, by May 2022, the patient was living with metastatic disease, showing features of advanced ductal carcinoma with a higher Ki67 index (20%) and extracellular mucin. Unfortunately, next-generation sequencing failed to identify any actionable genetic targets, and the patient transitioned to palliative care.
GeneSilico’s Retrospective Predictive Insights
Looking back, GeneSilico’s AI-powered predictive model was put to the test. Using the patient’s bulk transcriptomics data, GeneSilico accurately predicted several chemotherapy agents that would likely yield positive outcomes. Among the drugs identified, Cyclophosphamide and Paclitaxel stood out—both of which played pivotal roles in the patient's real-life treatment, resulting in a complete clinical response. This correlation between GeneSilico’s predictions and the patient's initial treatment supports the model’s capacity to guide treatment decisions in complex cases, even with limited data.
Additionally, GeneSilico’s predictions flagged several agents that would have been less effective for this patient, including Gemcitabine, Ixabepilone, Eribulin, Capecitabine, and Methotrexate. These insights could have potentially steered clinicians away from less promising treatment avenues, preventing unnecessary toxicity or ineffective treatments.
GeneSilico’s exome profiling also uncovered a significant finding—a somatic CBR3 V244M mutation that predisposes patients to Doxorubicin-induced cardiotoxicity. While Doxorubicin was part of the patient’s treatment (as FAC), this mutation might have helped clinicians balance the risks and benefits more precisely had it been known earlier, mitigating the potential for severe side effects.
Importance of Serial Molecular Profiling
It is also noteworthy that the tissue block used for GeneSilico’s transcriptomic and exome sequencing was collected during the patient's initial surgery. While this provided valuable insights early in the treatment journey, serial profiling of subsequent tissue samples could have improved the chances of detecting actionable therapeutic options as the cancer evolved. As the disease progressed, ongoing molecular profiling may have revealed additional mutations or drug sensitivities, enabling more precise and adaptable treatment strategies.
Conclusion
This case highlights the predictive power of GeneSilico’s AI-driven platform in precision oncology. By analyzing both transcriptomics and exome data, GeneSilico was able to align its predictions with the patient’s real-world treatment outcomes. These findings emphasize the platform’s ability to not only identify effective therapies but also flag potential risks, offering an invaluable tool for guiding clinical decisions in even the most challenging metastatic cancer cases. The case also underscores the potential of serial profiling in optimizing precision oncology strategies through ongoing molecular monitoring.