MRI PCa Screening | FDA AI Bias | Angio+MRI

“Whether it’s AI or traditional models, we have always been complacent . . .”

MGH’s Connie Lehman, MD, PhD on breast imaging’s biased history.

Imaging Wire Sponsors

  • Arterys – Reinventing imaging so you can practice better and faster.
  • Bayer Radiology – Providing a portfolio of radiology products, solutions, and services that enable radiologists to get the clear answers they need.
  • GE Healthcare – Enabling clinicians to make faster, more informed decisions through intelligent devices, data analytics, applications and services.
  • Healthcare Administrative Partners – Empowering radiology groups through expert revenue cycle management, clinical analytics, practice support, and specialized coding.
  • Hitachi Healthcare Americas – Delivering best in class medical imaging technologies and value-based reporting.
  • Novarad – Transformational imaging technologies that empower hospitals and clinicians to deliver clinical, operational and fiscal excellence.
  • Nuance – AI and cloud-powered technology solutions to help radiologists stay focused, move quickly, and work smarter.
  • Riverain Technologies – Offering artificial intelligence tools dedicated to the early, efficient detection of lung disease.
  • Siemens Healthineers – Shaping the digital transformation of imaging to improve patient care.
  • Zebra Medical Vision – Transforming patient care with the power of AI.

The Imaging Wire

A Case for MRI Prostate Cancer Screening

A new study in JAMA Oncology highlighted Short MRI’s prostate cancer screening potential, suggesting that it could be a PSA testing alternative and a better option than multiparametric MRI.

  • The Study – The researchers performed biparametric “Short” MRI (T2 weighted and diffusion), transrectal ultrasound, and PSA prostate cancer screening tests on 408 men. They used a 5-point prostate cancer suspicion scale for the MRI/US exams and defined suspicious PSA tests as PSA ≥3 μg/L.
  • The 3-5 Results – When the researchers used 3-5 scores as their suspicion range, MRI and ultrasound led to far more biopsy recommendations than PSA (72 MRI, 96 US, 40 PSA) and MRI detected two-times more clinically significant cancers (14 MRI, 9 US, 7 PSA).
  • The 4-5 Results – When they used 4-5 scores as their suspicion range, MRI and ultrasound screenings led to a similar number of biopsy recommendations as PSA testing (43 MRI, 52 US, 40 PSA) and MRI still detected more clinically significant cancers than PSA (11 MRI, 4 US, 7 PSA).
  • The Takeaway – Using short non-contrast MRI and a 4-5 prostate cancer suspicion threshold could improve detection of clinically significant prostate cancer without increasing biopsy volumes or overdiagnosis. Short MRI could also be a better screening option than multiparametric MRI, which takes longer (30-40 min vs. 15 min) and requires contrast.
  • Up Next – The researchers (you guessed it) called for more research into MRI’s potential role in prostate cancer screening, specifically suggesting studies that evaluate different screening intervals, feature larger or randomized cohorts, and address interobserver variability.

The Wire

  • Breast AI’s FDA Disparities: A STAT News investigation suggests that the FDA’s policy to approve AI devices without requiring developers to disclose their validation data’s racial demographics could worsen breast cancer disparities. STAT found that only one of 10 FDA-cleared breast imaging AI products disclosed its validation set’s racial demographics, suggesting that the FDA is still figuring out how to balance public trust and commercial interests when it comes to dataset transparency.
  • ACR Wants Value-Based Recognition: The ACR posted a letter to CMS’ MIPS policy contractor, Acumen, proposing new measures that would recognize radiologists’ role in value-based care. The ACR specifically called for new episode-based MIPS measures that would recognize radiologists for breast cancer coordination (1yr from screening to diagnosis/follow-up; measured by CDRs, recalls, true/false positives) and incidental findings (ensuring appropriate follow-ups, avoiding unnecessary imaging).
  • Angio + MRI for AIS: New research out of UCSF found that integrating angiography and MRI into acute ischemic stroke (AIS) treatment helps guide revascularization decision making. The researchers observed AIS endovascular interventions performed on 12 patients using angiography + MRI and 35 patients using angiography-alone, finding that MRI influenced whether to: 1) Perform initial / additional mechanical thrombectomy; 2) Place an intracranial stent; 3) Administer antithrombotic or blood pressure medications.
  • The First AI CPT Code: Four of the major radiology societies (ACR, RSNA, ARRS, AUR) asked the AMA CPT Editorial Panel to create a CPT code for AI-based vertebral fracture detection, which (if approved) would become the first radiology-specific artificial intelligence CPT code in 2022.
  • Breast Imaging Class Action: Sudbury Health Sciences North hospital in Ontario, Canada is facing a class action lawsuit that claims “systemic problems” in its breast cancer imaging operations led to “‘countless missed lesions” and “near catastrophic outcomes.” The class action names one woman who is currently battling metastatic cancer after a missed diagnosis (and is suing for CA$22m), but also seeks compensation for all affected patients from 2008 to 2020. The case is supported by accounts from Sudbury’s own clinicians who detail declining breast imaging standards, repeated internal quality warnings, and little efforts to improve division performance.
  • Breast Cancer Mortality Flattens Under-40: A new study in RSNA revealed that breast cancer mortality rates among 20yr-39yr women stopped declining from 2010 to 2017 (20–29yr +2.8% per year, 30-39yr +0.3% per year; both “nonsignificant”), after consistent mortality rate declines from 1987 to 2009. That’s very different from 40yr-79yr women, who achieved mortality rate declines during both the 1987-2009 and 2010-2017 periods (2010-2017: -1.2% to -2.2% per year). The researchers attributed the younger women’s flattening mortality rates in-part to an increase in distant-stage breast cancer (+4% per year since 2000), while forecasting that mortality rates among 20-39yr women may start to increase in the coming years.
  • NHS Diagnosis Delays: COVID-related screening delays made 2020 the NHS’ worst-ever year for cancer diagnosis. England’s public healthcare system recorded 200,940 urgent cancer referrals in 2020 (vs. 187,811 in 2019) and 350k patients had to wait over six weeks for key diagnostic tests (vs. 42k in 2019).
  • Rhino Health’s $5m: Healthcare federated learning technology developer Rhino Health wrapped up a $5m Seed round that it will use to further develop its multi-institution/site AI development platform. Rhino Health is founded by a pair of MGH and Google AI specialists, it played a key role in a previous Nvidia federated learning imaging project, and it has a partnership with Nvidia to bring its federated learning solution to more hospitals.
  • LUS for COVID Assessments: A new study out of Brazil provided even more evidence of lung ultrasound’s value for COVID-19 assessments. In the prospective study, emergency physicians calculated lung ultrasound scores on 180 COVID-positive patients (12 lung regions, 0-3 scores per region, 0-36 max score), finding that ≥20 LUS scores were associated with ICU admission (20 avg. LUS), intubation (21.3 vs. 15.2 avg.), and mortality (21.6 vs. 16.7 avg.).
  • Novasignal Adds Communication & Progression: Novasignal added new features to its Novaguide cerebral ultrasound system (combines a robotic headset w/ doppler US) to support clinical communication and ongoing patient management. The new communication app directly alerts clinicians when Novaguide detects potentially serious changes in blood flow (e.g. stroke), while a new AI tool collects data from every scan to provide insights into patient progress.
  • RVUs ≠ Effort: A new Clinical Imaging study found that the current RVU system doesn’t match radiologists’ effort (especially with certain types of imaging), suggesting that a time-based metric could be more representative. A UPenn researcher evaluated 242k radiology reports and calculated study ascribable times (SATs) for each CPT code across all radiologists, revealing wide variations in the time that different radiologists would require to achieve the AAARAD 65th percentile. This variability was most notable for MSK imaging studies (141%, 198 days), followed by neuroimaging (107.5%, 162 days), body imaging (89.6%, 186 days), and thoracic imaging (72%, 84 days).
  • ClearPoint & Blackrock: MRI-based neurosurgery platform company ClearPoint Neuro announced a co-development alliance with neuroscience/neuro-electrode technology company Blackrock Microsystems that will bring ClearPoint Neuro “beyond the MRI suite and into the Operating Room.” The companies will first launch their co-developed Microelectric Recording platform in 2023, followed by “smart” biopsy needles and other implantable neural electrodes.

The Resource Wire

– This is sponsored content.

  • This Hitachi blog details how a complete CVIS solution drives efficiencies and increases productivity.
  • Watch this recorded webinar from Healthcare Administrative Partners where they examine how the 2021 Medicare Physician Fee Schedule (MPFS) and Quality Payment Program (QPP) final rules will impact radiologists.
  • This Riverain Technologies case study details how Einstein Medical Center adopted ClearRead CT enterprise-wide (all 13 CT scanners) and how the solution allowed Einstein radiologists to identify small nodules faster and more reliably.
  • Check out this GE Healthcare perspective on imaging AI’s adoption evolution, the AI ecosystem shift, and the importance of seamless AI integration.
  • With a post-COVID surge in orthopedic surgery looming, this Zebra Med blog details how its 3D imaging orthopedic solution can help surgeons improve their preoperative efficiencies.
  • Check out how Novarad’s CryptoChart image sharing solution allows patients to easily access and share their medical images using personalized, highly secure QR codes (rather than CDs).

You might also like

You might also like..

Select All

You're signed up!

It's great to have you as a reader. Check your inbox for a welcome email.

-- The Imaging Wire team

You're all set!