Cracking Open the Other Black Box: Unmasking AI's Role in Patient Care

Oct 25, 2023

Let’s talk breast cancer screening, access to results in the portal, and AI powered breast cancer detection.

It was exciting to see my local imaging center offering Enhanced Breast Cancer Detection (EBCD) using Deep Health’s SAIGE-Dx, a mammography diagnostic aid software.

As someone who has a family history of breast cancer and extremely dense breasts, I’ve been getting routine mammograms with ultrasounds since my early 30’s. About 10% of women fall into the category of dense breasts, where it is extremely difficult to detect a tumor on a mammogram. Think trying to find a snowball in a blizzard. If your life depended on finding that snowball in a blizzard, wouldn’t you want a tool that could improve your odds of finding it?

Intrigued, I invested in the $40 out-of-pocket expense for EBCD as an add-on to my mammogram.

When my mammogram results hit my portal, I stopped dead in my tracks seeing “Saige-Dx, an Artificial Intelligence Diagnostic Aid Software” listed in the technique description; see screenshot below. So simple yet so powerful. As a patient and advocate, it’s refreshing to have transparency into AI being used as part of my care. While AI-powered clinical decision support may be routinely used inside the 4 walls of medicine, it’s often not mentioned in the clinical notes or in the medical record that’s accessible to patients.

EXAM: DIGITAL BILATERAL SCREENING MAMMOGRAM AND TOMOSYNTHESIS WITH CAD AND BREAST ULTRASOUND

MAMMOGRAM:

HISTORY: Female, 42 years old, presents for screening mammogram. No personal history of breast cancer. Family history of breast cancer in her mother diagnosed at age 48.

TECHNIQUE: Biltaeral 3D digital screening mammogram was performed. Examination was reviewed with Saige-Dx, an Artificial Intelligence Diagnostic Aid Software. Synthesized 2D images (C-view) were reconstructed from digital breast tomosynthesis data.

Thankfully my mammogram results were normal but there were a few benign findings noted that had never been reported before, including benign calcifications and tiny cysts. Were these detected by Saige-Dx or by the radiologist, or a combination of both? Had these benign findings always been there and only now newly detected due to the sensitivity of the EBCD service? Were these genuinely new findings thanks to 43 being around the corner? I needed to know more.

I called the imaging place to request a copy of my mammogram report WITHOUT Saige-Dx’s EBCD software analysis. Full disclosure, I was very clear to emphasize how excited I was that my imaging center was offering this add on service. I was not calling to complain. I was calling because I was curious. I was connected to the office manager who seemed perplexed by my request and said they had to speak to a radiologist and get back to me.

I want a side-by-side comparison: my mammogram with and without the AI Enhanced Breast Cancer Detection. I paid $334.90 for my mammogram (thanks to still having not met my high OOP annual deductible) plus $40 for my EBCD. I want the results I paid for and have my questions answered. After numerous phone calls and requests, it’s clear that the imaging center doesn’t know what to do with my request and is leaning into radio-silence. I’m not giving up; stay tuned.

There are 2 major takeaways that are important to highlight:

Patients have a right to know about the role that AI may play in their care, from point of diagnosis, to guiding treatment planning, to clinical trial selection, to reimbursement and payment, as well as coordination of care. The way we disseminate radiology results is well overdue for an overhaul. We must be future-facing in how we disseminate results powered by AI. We need the patient stakeholder as a driver of this co-creation process to ensure results are easy to understand and actionable.
I fully recognize my privilege of being able to afford the mammogram as well as the out-of-pocket cost for the EBCD service. As more consumer/patient-facing AI powered innovations hit point of care, how can we ensure equity in access and address financial aid and patient assistance programs?

Cracking Open the Other Black Box: Unmasking AI's Role in Patient Care

Related Article

Nightmare on Prior Auth Street