A 20-Year Survivor Wonders: Could AI Have Personalized His Bladder Cancer Treatment in 2005?
A physician reflects on his 2005 bladder cancer journey, questioning whether the emerging power of artificial intelligence could have delivered a more tailored – and potentially life-saving – treatment plan than was available at the time.
The fall of 2005 began with a startling sight: a single drop of blood in the toilet bowl after urination. Described as a “rose-colored bead” that appeared intermittently over several weeks, it initially didn’t raise alarm. He was a relatively young, healthy man with no notable medical history beyond mild, well-controlled hypertension, and no history of obesity, or high blood pressure. However, a family history of urogenital cancers in his adoptive parents, coupled with his status as an adoptee with no knowledge of his biological lineage, raised concerns about potential genetic predispositions or environmental factors.
The adage “painless hematuria is cancer untill proven otherwise” resonated with increasing urgency as the episodes continued. An ultrasound revealed a “soft tissue density” in his bladder, prompting a cystoscopy. The diagnosis was swift and sobering: a high-grade urothelial carcinoma, extensively invading the bladder wall with evidence of aggressive spread. At the time, the five-year survival rate for stage II muscle-invasive bladder cancer hovered around 45 percent.
Faced with a cancer typically affecting those decades older, the physician felt a sense of urgency. Despite the grim prognosis,he possessed advantages – access to leading Boston-based medical centers,specialist colleagues,and comprehensive insurance. However, the path forward was fraught with uncertainty. Three urologists uniformly recommended a radical cystectomy, involving bladder removal and reconstruction. This convergence of opinion offered a degree of certainty regarding the surgical approach.
The challenge lay in determining the optimal chemotherapy regimen. In the mid-2000s, roughly 500,000 new research publications were being indexed annually on PubMed, creating an overwhelming volume of information for oncologists. Treatment decisions relied on NCCN/ASCO guidelines, randomized controlled trials, and meta-analyses. Yet, the physician encountered conflicting recommendations from three medical oncologists regarding the intensity and timing of chemotherapy. “The wolf is already out of the cage,” one oncologist warned, highlighting the likelihood of microscopic disease beyond the bladder.
Faced with a “dartboard toss” of treatment options, the physician ultimately relied on intuition. He chose a regimen based on his gut feeling, hoping for the best. A subsequent FISH analysis revealed a small subclone of HER2 amplified cells in his cancer. Based on the success of trastuzumab (Herceptin) in HER2-positive breast cancer,his oncologist consulted with a colleague at the University of michigan specializing in HER2 and bladder cancer.He agreed to add Herceptin to his regimen, driven by a desire for a survival advantage.
Twenty years later, the physician is a grateful survivor. However, he now reflects on how far cancer care has evolved. The “nuclear bomb” approach of the past is giving way to a more precise “stealth bomber” strategy, fueled by advancements in next-generation sequencing (NGS), liquid biopsies (ctDNA and cfDNA assays), CAR-T cell therapy, and other cutting-edge technologies.
While these tools are powerful, their effective request requires discerning which ones are best suited for an individual patient’s unique cancer characteristics. This is where artificial intelligence (AI) is poised to revolutionize oncology.Driven by large language models, AI can synthesize vast amounts of data, predict treatment responses, and personalize recommendations with unprecedented accuracy.It promises to move beyond pattern recognition and guidelines,offering “fine scalpels,not blunt instruments” guided by iterative learning.
The physician wonders what a data-driven personalization platform would have recommended in his “anomalous N-of-1 situation” back in 2005. While the answer remains unknown, he expresses optimism about the future of cancer care, believing that customized treatments will lead to better patient outcomes. He emphasizes, though, that early detection remains paramount. “Hope lives there, too.”
George Beauregard, DO is an Internal Medicine physician & the author of Reservations for nine: A Doctor’s Family Confronts Cancer.
