George Beauregard
How Artificial Intelligence Personalized My Cancer Journey in 2005
I don’t think I’m in the minority among baby boomer doctors when it comes to my curiosity and ambivalence about the progressive applications of artificial intelligence in medicine. But this curiosity is not only prospective, but also retrospective. In 2005, I became an outsider who might need something more than standard care for my disease.
In the fall of 2005, I first saw a drop of blood in the toilet bowl while I was peeing in the bathroom. After falling into the water, the rose-colored beads slowly sank, twisted and twisted, and dissipated like a wisp of smoke. Evidence is ephemeral—gone in seconds. If I were a spectator rather than a source, I might appreciate its visual artistry. There is no associated pain.
A thought flashed through my mind: Was the blood I peed just now? I thought maybe I was imagining it.
I was 49 years old and had no risk factors considered to be kidney or bladder cancer: smoking, obesity, advanced age, high blood pressure, or exposure to cadmium, trichlorethylene, or herbicides. But I was adopted and knew nothing about my family history. Do I have a grim family tree? Perhaps important, however, is that both my adoptive parents had different types of genitourinary cancer. This leads me to speculate that environmental factors related to the materials in our house and/or the land it is on or surrounding, may have played a role.
I tried to allay any worries but maxim “Painless hematuria is cancer unless proven otherwise” went through my head in a chyron-like way.
The condition continued to worsen and an ultrasound was required, the report states: “…Soft tissue density was seen at the base of the right bladder. Although this may represent a thrombus, I cannot rule out a primary mucosal lesion. The diameter of the lesion is approximately 4 X 5 cm. “
I consulted a urologist colleague and he performed a cystoscopy. His comment on what he saw was: “As you know, you have a mass in your bladder. I saw it clearly. It looked angry, so I suspected it was not benign. I tried to remove as much of it as I could. If it was scraped deeper, it could puncture your bladder, which is quite dangerous. I know I didn't fully understand it.” The TURBT was quickly performed. Pathology showed high-grade urothelial carcinoma with extensive invasion of the lamina propria and muscularis propria. There is multifocal lymphovascular invasion, so my subgroup may be more advanced than local SEER stage.
At that time, the relative five-year survival rate for stage II muscle-invasive bladder cancer was about 45 percent.
Most often, bladder cancer is an age-related malignancy. So, I am 49 years old and have cancer whose median age of onset (seventies) is much older than mine. WTF moment.
It got me thinking about how much time I had left.
So, I have cancer, but in some ways I feel cautiously optimistic. I had access to an academic center in Boston and expert colleagues willing to see me quickly, and with good insurance.
But getting a diagnosis is just the beginning. I saw three urologists who all recommended radical cystectomy, small bowel resection, and orthotopic ileal neobladder construction. convergence. Certainty for me.
In the mid-2000s, approximately half a million new research publications were indexed in PubMed. At that time, oncologists would typically begin research on complex cases based on NCCN/ASCO guidelines (synthesis of evidence), examine supporting randomized controlled trials (gold standard), meta-analyses, and possibly consult ClinicalTrials.gov to learn about new or ongoing studies before making treatment recommendations.
I also met with three medical oncologists from different prestigious academic medical centers. One of the memorable comments was: “The wolf is out of the cage,” implying that the potential for widespread microscopic disease beyond the bladder was high.
Each of them recommended regimens that were known and available at the time: different “seventy-year-old-friendly” chemotherapy regimens, including the type and number of drugs used (double, triple, quadruple) and the timing of their administration relative to surgery (neoadjuvant, adjuvant, or half and half). Opinions are conflicting. Disagreement. Uncertainty for me.
The lack of conclusive evidence on which treatment regimen confers a longer survival benefit feels like a dart toss to me. I wonder if my choices will keep me underwater, but ultimately be able to surface instead of drowning. My decision-making process ends up being driven mostly by intuition. I told myself, make a choice and don’t look back.
In 2005, the benefits of adding trastuzumab (Herceptin) to treat HER-2-positive breast cancer were confirmed. The oncologist I chose spoke with a colleague at the University of Michigan, a researcher who focuses on HER2 and bladder cancer. The FISH data of my cancer cells show a subclones of HER2-amplified cells; the percentage is indeterminate but low. After discussing the harm-to-benefit ratio of adding Herceptin to my regimen, I agreed. For me, this decision was less about satisfying academic curiosity and more about existential advantage.
So, here I am, and, for the most part, I’m a grateful (and I think lucky) survivor of 20 years.
But how has the field of oncology changed since then, as cancer treatment has gradually shifted from the old universal nuclear bomb approach to the stealth bomber approach.
Today's oncologists' black bags contain new and enhanced tools at their disposal. Enhancements in NGS, ctDNA and cfDNA detection, CAR-T cell therapy, qPCR and RT-PCR, spatial transcriptomics, epigenetic analysis techniques, mass spectrometry-based proteomics, epigenetic analysis techniques, and more. Medical Frontiers.
While it's nice to have many more sophisticated tools, if a diagnostician or prosthetist doesn't know exactly which one works best for an individual person's unique combination of cancer characteristics, they'll be digging back into the medical literature, remembering what has worked (or worked to some degree), pattern recognition, guidance, and intuition for other “similar” patients.
In the pursuit of precision medicine, artificial intelligence, a powerful ally, is accelerating from a sidecar to a prime mover, driven by large-scale language models that can collect, assimilate and organize hitherto unimaginable numbers of disparate, clinically meaningful data points and synthesize them to predict and guide treatment options, avoid unseen and unforeseen dead ends and rabbit holes in the future, and tailor treatment recommendations to individual patients. And make course corrections as needed along the way. Decipher medical codes. At warp speed.
A fine scalpel, not a blunt instrument, guided by iterative learning and adaptation.
While I'm glad I'm still here, I wonder what a data-driven personalization platform would recommend for my unusual N-of-1 situation.
I will never know, but my optimism and hope for more results in the future of the effectiveness of individually tailored cancer care are growing. While it will never be perfect, it could mean better patient outcomes will be achieved.
One important thing remains: meaningful cancer detection as early as possible at lower stages. Hope exists here, too.
George Beauregard, DO is an internist with more than 20 years of experience in clinical practice and leading organizational strategic and clinical initiatives. This comes from his substack
