Author: Brian Casey

AI’s Incremental Revolution

Posted on by Brian Casey

So AI dominated the discussion at last week’s RSNA 2023 meeting. But does that mean it’s finally on the path to widespread clinical use? 

Maybe not so much. For a technology that’s supposed to have a revolutionary impact on medicine, AI is taking a frustratingly long time to arrive. 

Indeed, there was plenty of skepticism about AI in the halls of McCormick Place last week. (For two interesting looks at AI at RSNA 2023, also see Hugh Harvey, MD’s list of takeaways in a post on X/Twitter and Herman Oosterwijk’s post on LinkedIn.) 

But as one executive we talked to pointed out, AI’s advance to routine clinical use in radiology is likely to be more incremental than all at once. 

  • And from that perspective, last week’s RSNA meeting was undoubtedly positive for AI. Scientific sessions were full of talks on practical clinical applications of AI, from breast AI to CT lung screening. 

Researchers also discussed the use of AI apart from image interpretation, with generative AI and large language models taking on tasks from answering patient questions about their reports to helping radiologists with dictation.

It’s fine to be a skeptic (especially when it comes to things you hear at RSNA), but for perspective look at many of the past arguments casting doubt on AI: 

  • AI algorithms don’t have FDA clearance (the FDA authorized 171 algorithms in just the past year)
  • You can’t get paid for using AI clinically (16 algorithms have CPT codes, with more on the way) 
  • There isn’t enough clinical evidence backing the use of AI (tell that to the authors of MASAI, PERFORMS, and a number of other recent studies with positive findings)
  • The AI market is overcrowded with companies and ripe for consolidation (what exciting new growth market isn’t?)

The Takeaway

Sure, it’s taking longer than expected for AI to take hold in radiology. But last week’s conference showed that AI’s incremental revolution is not only advancing but expanding in ways no one expected when IBM Watson was unveiled to an RSNA audience a mere 6-7 years ago. One can only imagine what the field will look like at RSNA 2030.

Looking for more coverage of RSNA 2023? Be sure to check out our videos from the technical exhibit floor, which you can find on our new Shows page.

AI Dominates at RSNA 2023

Posted on by Brian Casey

Take a deep breath. You survived another RSNA conference.

While a few hardy souls are still enjoying educational sessions in the cozy confines of McCormick Place, the final day of the exhibit floor yesterday marks the end of RSNA 2023 for most attendees. And what a show it was. 

Predictions were that AI would dominate the scientific sessions at RSNA 2023, a forecast that largely panned out. A November 28 session was a case in point, in which a series of top-quality papers were presented on one of the most promising use cases of AI, for breast screening:

  • A homegrown AI algorithm that analyzed screening breast ultrasound exams in addition to FFDM and DBT mammograms boosted sensitivity for detecting cancer in 12.5k patients, with better sensitivity for women with dense breasts (71% vs. 60%) and non-dense breasts (79% vs. 63%)
  • AI did a good job of detecting breast arterial calcification (BAC) when used prospectively to analyze screening mammograms in 16k women across 15 sites.  It found 15% of women had BAC, a possible marker for atherosclerotic disease
  • Swedish researchers used their VAI-B validation platform to compare three AI algorithms (Therapixel, Lunit, and Vara) in 34k women, finding that using AI with a single radiologist boosted sensitivity 10-30% compared to double reading, with a slight loss in specificity (2-7%). VAI-B could be used to validate AI implementation and guide purchasing decisions
  • Why does AI miss some breast cancers? South Korean researchers addressed this question by analyzing 1.1k patients with invasive cancers in which AI had a miss rate of 14%. Luminal cancers were missed most often
  • Adding AI analysis of prior images to current studies with FFDM and DBT boosted sensitivity for cancer detection in 30k patients, with sensitivity the highest for two years of priors compared to no priors (74% vs. 70%)

The Takeaway

This week’s research points to an exciting near-term future in which AI will help make mammography screening more accurate while helping breast radiologists perform their jobs more efficiently. Landmark studies toward this end were published in 2023 – this week’s RSNA conference shows that we can expect the momentum to continue in 2024. 

Welcome to RSNA 2023

Posted on by Brian Casey

It’s off to the races at RSNA 2023 as radiology’s showcase conference kicked off on Sunday. 

“Leading Through Change” is the theme of this year’s meeting, and it’s an appropriate slogan for a specialty that seems on the cusp of disruption with the growing use of AI, deep learning, and other tools. 

  • AI is being featured prominently in scientific presentations and vendor exhibits in McCormick Place, with a particular focus on whether large language models like ChatGPT can find practical application in radiology. Early research is promising but still inconclusive.

Another major focus at RSNA 2023 has been lung cancer screening, with Sunday afternoon sessions investigating how screening can be expanded: 

  • Researchers mined a database of 32k women who got screening mammography to find eligible candidates for lung screening, finding 5% who met screening criteria. 
  • Using the USPTSF’s 2021 guideline revision to find screening candidates led to shorter smoking histories (42 vs. 29 pack-years) and slightly more women being eligible (48% vs. 46%). 
  • ChatGPT gave more correct answers than Google Bard to non-expert questions on lung screening (71% vs. 52%).
  • ChatGPT, GPT-4, and Bard needed multiple iterations to produce reports readable by patients. 

AI is also proving its value for selecting screening candidates and identifying lung pathology: 

  • An AI algorithm analyzed chest X-rays to determine whether an individual would benefit from CT lung cancer screening – even if they don’t smoke. In 17.4k patients, the model classified 28% as high risk, 2.9% of whom were later diagnosed with lung cancer, a higher level than the 1.3% six-year threshold at which guidelines recommend CT lung screening.
  • A deep learning algorithm analyzed chest X-rays in a cohort of 10k patients to predict who would develop type 2 diabetes, turning in better accuracy than a model that only looked at clinical factors like age, BMI and HbA1c levels (AUCs:  0.84 vs. 0.79). 

Looking for more coverage of RSNA 2023? Be sure to check out our videos from the technical exhibit floor, which you can find on our new Shows page. 

The Takeaway
The RSNA has always been known as the Super Bowl of radiology, and this year’s meeting is off to a great start. Be sure to check back on our Twitter/X, LinkedIn, and YouTube pages for more coverage of this week’s events in Chicago.

Vendors Enter RSNA on Q3 Roll

Posted on by Brian Casey

As RSNA 2023 approaches, medical imaging vendors appear to be on a roll when it comes to financial results. In the weeks leading up to the meeting, companies have posted numbers that for the most part are strongly positive and appear to be leaving the bad old days of the COVID-19 pandemic behind.

Agfa – Between Agfa’s two imaging divisions, healthcare IT continues to outperform the radiology solutions business. Healthcare IT saw growth in revenue (3.3% to $67M) and EBITDA (44.3% to $6.4M), but revenue declined at radiology solutions (-5.7% to $127M) as did EBITDA (-21% to $10M). 

Canon – Canon Medical Systems saw firm revenues in Japan and Europe, which propelled the business unit to higher revenues (5% to $913M) while income before taxes edged up (0.3% to $46M). 

Fujifilm – Revenues tapered off slightly in Fujifilm’s healthcare business at constant currency rates (-1.9% to $1.66B) as a 12.4% decline in its contract manufacturing business offset 1.7% growth in medical systems. Operating income in healthcare slipped due to a one-time benefit in the year-ago quarter (-6.5% to $217M).

GE HealthCare – Revenue growth in its molecular imaging and CT businesses helped propel GE HealthCare’s revenue growth (5.4% to $4.82B), assisted by 13% growth in pharmaceutical diagnostics and a 9% increase in patient care solutions. Net income was lower (-23% to $375M). 

Guerbet – Strong revenues for the third quarter in Asia (+15%) and stability in the EMEA region (0.6%) helped counter a decline in the Americas (-5.2%), enabling Guerbet to turn in overall quarterly revenue growth at constant exchange rates (2.3% to $212M). The company expects sales of its Elucirem MRI contrast agent to ramp up in the fourth quarter. 

Hologic – The semiconductor shortage that had impacted Hologic in previous quarters eased, leading to a sharp jump in revenues in the company’s breast health business (27% to $353M). The rebound didn’t extend to Hologic’s overall net income as its net margin narrowed (-24% to $91M). 

Konica Minolta – A decline in sales of X-ray systems to hospitals in its core market of Japan and a slower US hospital market produced lower revenues in Konica Minolta’s healthcare division (-5% to $238M), and the business posted an operating loss (-$5.5M).

Philips – Philips rebounded in the most recent quarter, with revenues in its diagnosis and treatment division rising sharply after currency conversion thanks to double-digit growth in all businesses (14% to $2.39B). Operating income doubled (to $272M). 

RadNet – RadNet saw a double-digit jump in revenues (15% to $402M) while net income leaped ($17.5M vs. $668k). Revenue jumped 221% in the company’s AI segment, which made progress narrowing its EBITDA loss (-$2.5M vs. -$4.5M) on higher consumer adoption of its Enhanced Breast Cancer Detection offering.  

Siemens Healthineers – Siemens Healthineers closed its financial year with “outstanding” 8.3% revenue growth at constant exchange rates, including double-digit growth in its imaging business (11% to $3.62B) while adjusted EBIT edged up (2% to $812M). Its Varian radiation therapy business saw a strong recovery in revenue (30% to $1.1B) and adjusted EBIT (90% to $207M).

Varex – Growth in Varex’s industrial X-ray imaging business propelled the company to higher overall revenues even as revenues in its medical business fell (-9.8% to $164M). The medical division’s gross profit also slipped (-7% to $53M).

The Takeaway

Not every company was a winner in this last round of quarterly earnings, but at least the macroeconomic headwinds of the COVID-19 pandemic are fading. The fourth calendar quarter is typically radiology’s strongest period due to the impact of the RSNA conference on equipment purchasing, so let’s hope the momentum continues.

Reimbursement Drives AI Adoption

Posted on by Brian Casey

It’s no secret that insurance reimbursement drives adoption of new medical technology. But a new analysis in NEJM AI shows exactly how reimbursement is affecting the diffusion into clinical practice of perhaps the newest medical technology – artificial intelligence. 

Researchers analyzed a database of over 11B CPT claims from January 2018 to June 2023 to find out how often reimbursement claims are being submitted for the use of the over 500 AI devices that had been approved by the FDA at the time the paper was finalized. 

  • The authors chose to focus on CPT claims rather than claims under the NTAP program for new technologies because CPT codes are used by both public and private payors in inpatient and outpatient settings, while NTAP only applies to Medicare inpatient payments. 

They found 16 medical AI procedures billable under CPT codes; of these, 15 codes were created since 2021 and the median age of a CPT code was about 374 days, indicating the novelty of medical AI.

  • Also, only four of the 16 had more than 1k claims submitted, leading the authors to state “overall utilization of medical AI products is still limited and focused on a few leading procedures,” such as coronary artery disease and diabetic retinopathy.

The top 10 AI products and number of CPT claims submitted are as follows:

  1. HeartFlow Analysis for coronary artery disease (67,306)
  2. LumineticsCore for diabetic retinopathy (15,097)
  3. Cleerly for coronary atherosclerosis (4,459)
  4. Perspectum LiverMultiScan for liver MRI (2,428)
  5. Perspectum CoverScan for multiorgan MRI (591)
  6. Koios DS for breast ultrasound (552)
  7. Anumana for ECG cardiac dysfunction (435)
  8. CADScor for cardiac acoustic waveform recording (331)
  9. Perspectum MRCP for quantitative MR cholangiopancreatography (237)
  10. CompuFlo for epidural infusion (67)

While radiology may rule in terms of the sheer number of FDA-approved AI products (79% in a recent analysis), the list shows that cardiology is king when it comes to paying the bills. 

The Takeaway

Amid the breathless hype around medical AI, the NEJM AI study comes as a bit of a wake-up call, showing how the cold reality of healthcare economics can limit technology diffusion – a finding also indicated in other studies of economic barriers to AI. 

On the positive side, it shows that a rosy future lies ahead for those AI algorithms – like HeartFlow Analysis – that can make the leap.

Uneven Success Against Breast Cancer

Posted on by Brian Casey

The decline in breast cancer mortality has been one of public health’s major success stories. But when you look at it from a global perspective, it’s the best of times and the worst of times. 

That’s because success in fighting breast cancer has been uneven around the world. While countries in North America, Western Europe, and Oceania have seen dramatic declines in breast cancer mortality and advanced-stage disease, other regions continue to be plagued by what really is becoming a survivable disease for most women. 

A new study in JAMA Oncology points out these disparities, documenting major differences in rates of advanced breast disease between countries in what researchers said was the most comprehensive review to date of global differences in breast cancer stage at diagnosis. 

  • Researchers conducted a meta-analysis of 133 studies covering 2.4M women across 81 nations over the past two decades, documenting differences in rates of advanced breast disease at diagnosis both over time and between countries. 

While most high-income nations have seen declines in rates of distant metastatic disease over the past 20 years, advanced-stage disease remains stubbornly common in lower middle-income countries. Researchers found: 

  • Rates of distant metastatic disease varied across countries by region, with sub-Saharan Africa the highest and North America the lowest (6-31% vs. 0-6%)
  • Lower socioeconomic status was tied to more advanced disease when women in the most disadvantaged group were compared to least disadvantaged (3-11% vs. 2-8%)
  • There were pronounced disparities even in high-resource countries with established screening programs, as rates of metastatic disease were twice as high in women of low socioeconomic status (SES) compared to high SES women, such as in the US (8% vs. 4%) 
  • Older women had a much higher prevalence of advanced disease across different countries compared to younger women (range of 4-34% vs. 2-16%), a phenomenon that could be because most screening programs stop at age 75
  • 40% of countries did not meet the Global Breast Cancer Initiative goal of having 60% or more of patients diagnosed at stage I or II

The Takeaway

The new findings indicate that it’s too soon to take a victory lap in the battle against breast cancer. While progress at higher socioeconomic levels in high-income countries has been impressive, breast cancer remains a scourge among more disadvantaged women and across wide regions of the world.

Lung Screening’s Long-Term Benefits

Posted on by Brian Casey

CT lung cancer screening produced lung cancer-specific survival over 80% in the most recent data from the landmark I-ELCAP study, a remarkable testament to the effectiveness of screening. 

The findings were published this week in Radiology from I-ELCAP, one of the first large-scale CT lung screening trials, and are the latest in a series of studies pointing to lung screening’s benefits. The findings were originally presented at RSNA 2022. 

The I-ELCAP study is ongoing and has enrolled 89k participants at over 80 sites worldwide from 1992-2022 who have been exposed to tobacco smoke and who received annual low-dose CT (≤ 3mGy) scans. Periodic I-ELCAP follow-up studies have documented the survival rates of those whose cancers were detected with LDCT, and the new numbers offer a 20-year follow-up, finding: 

  • Primary lung cancers were detected on LDCT in 1,257 individuals who had lung cancer-specific survival of 81%, matching the 10-year survival rate of 81%
  • 1,017 patients with clinical stage I lung cancer underwent surgical resection and saw a lung cancer-specific survival rate of 87%
  • The I-ELCAP survival rate is much higher than another landmark screening study, NLST, in which it was 73% for stage I cancer at 10 years
  • Lung cancer-specific survival hit a plateau after 10 years of follow-up, at a cure rate of about 80%

I-ELCAP is unique for a variety of reasons, one of which is that it continues to screen people beyond a baseline scan and 2-3 annual follow-up rounds – perhaps the reason for its higher survival rate relative to NLST. 

  • It also has included people who were exposed to tobacco smoke but who weren’t necessarily smokers – an important distinction in the debate over how broad to expand lung screening criteria.  

The findings come as CT lung cancer screening is generating growing momentum. Studies this year from Germany, Taiwan, and Hungary have demonstrated screening’s value, and several countries are ramping up national population-based screening programs. 

The Takeaway

The 20-year I-ELCAP data show that CT lung cancer screening works if you can get people to do it. But achieving survival rates over 80% also requires work on the part of healthcare providers, in terms of defined protocols for working up findings, data management for screening programs, and patient outreach to ensure adherence to annual screening. Fortunately, I-ELCAP offers a model for how it’s done.

More Support for CT Lung Cancer Screening

Posted on by Brian Casey

Yet another study supporting CT lung cancer screening has been published, adding to a growing body of evidence that population-based CT screening programs will be effective in reducing lung cancer deaths. 

The new study comes from European Radiology, where researchers from Hungary describe findings from HUNCHEST-II, a population-based program that screened 4.2k high-risk people at 18 institutions. 

  • Screening criteria were largely similar to other studies: people between the ages of 50 and 75 who were current or former smokers with at least 25 pack-year histories. Former smokers had quit within the last 15 years. 

Recruitment for HUNCHEST-II took place from September 2019 to January 2022. Participants received a baseline low-dose CT (LDCT) scan, with the study protocol calling for annual follow-up scans (more on this later). Researchers found: 

  • The prevalence of baseline screening exams positive for lung cancer was 4.1%, comparable to the NELSON trial (2.3%) but much lower than the NLST (27%)
  • 1.8% of participants were diagnosed with lung cancer throughout screening rounds
  • 1.5% of participants had their cancer found with the baseline exam
  • Positive predictive value was 58%, at the high end of population-based lung screening programs
  • 79% of screen-detected cancers were early stage, making them well-suited for treatment
  • False-positive rate was 42%, a figure the authors said was “concerning”

Taking a deeper dive into the data produces interesting revelations. Overdiagnosis is a major concern with any screening test; it was a particular problem with NLST but was lower with HUNCHEST-II. 

  • Researchers said they used a volume-based nodule evaluation protocol, which reduced the false-positive rate compared to the nodule diameter-based approach in NLST.

Also, a high attrition rate occurred between the baseline scan and annual screening rounds, with only 12% of individuals with negative baseline LDCT results going on to follow-up screening (although the COVID-19 pandemic may have affected these results). 

The Takeaway

The HUNCHEST-II results add to the growing momentum in favor of national population-based CT lung screening programs. Germany is planning to implement a program in early 2024, and Taiwan is moving in the same direction. The question is, does the US need to step up its game as screening compliance rates remain low?

Unpacking the Biden Administration’s New AI Order

Posted on by Brian Casey

It seems like watershed moments in AI are happening on a weekly basis now. This time, the big news is the Biden Administration’s sweeping executive order that directs federal regulation of AI across multiple industries – including healthcare. 

The order comes as AI is becoming a clinical reality for many applications. 

  • The number of AI algorithms cleared by the FDA has been surging, and clinicians – particularly radiologists – are getting access to new tools on an almost daily basis.

But AI’s rapid growth – and in particular the rise of generative AI technologies like ChatGPT – have raised questions about its future impact on patient care and whether the FDA’s existing regulatory structure is suitable for such a new technology. 

The executive order appears to be an effort to get ahead of these trends. When it comes to healthcare, its major elements are summarized in a succinct analysis of the plan by Health Law Advisor. In short, the order: 

  • Calls on HHS to work with the VA and Department of Defense to create an HHS task force on AI within 90 days
  • Requires the task force to develop a strategic plan within a year that could include regulatory action regarding the deployment and use of AI for applications such as healthcare delivery, research, and drug and device safety
  • Orders HHS to develop a strategy within 180 days to determine if AI-enabled technologies in healthcare “maintain appropriate levels of quality” – basically, a review of the FDA’s authorization process
  • Requires HHS to set up an AI safety program within a year, in conjunction with patient safety organizations
  • Tells HHS to develop a strategy for regulating AI in drug development

Most analysts are viewing the executive order as the Biden Administration’s attempt to manage both risk and opportunity. 

  • The risk is that AI developers lose control of the technology, with consequences such as patients potentially harmed by inaccurate AI. The opportunity is for the US to become a leader in AI development by developing a long-term AI strategy. 

The Takeaway

The question is whether an industry that’s as fast-moving as AI – with headlines changing by the week – will lend itself to the sort of centralized long-term planning envisioned in the Biden Administration’s executive order. Time will tell.

Canada’s Breast Screening Push

Posted on by Brian Casey

As Canada examines revisions to its breast cancer screening guidelines, a new study adds support to the proposal of lowering its screening age to 40 – a move made in the US earlier this year. 

When to start breast screening has long been one of the most controversial aspects of mammography. 

  • In the US, a firestorm erupted in 2009 when the USPSTF withdrew its recommendation that women start in their 40s … a policy that wasn’t rescinded until May. 

In Canada, the Canadian Task Force on Preventive Health Care is reviewing its 2018 screening guidelines, which currently advise women to wait until 50 to start routine breast screening, and then be screened every 2-3 years after that. 

  • The Canadian task force’s 2018 guidelines also don’t mention dense breast tissue, a known risk factor for breast cancer (the FDA earlier this year said it would begin requiring breast density reporting). 

Canadian breast specialists have been pushing for the task force to lower the screening age, and their efforts got a boost with a new study that found starting breast screening at age 40 and continuing with it annually saved the greatest number of lives.

Researchers in MDPI used the OncoSim-Breast microsimulation model to simulate various screening regimens in a cohort of 1.5M Canadian women born in 1975. They assessed the earlier screening strategy by various metrics, including impact on breast cancer mortality, number needed to be screened to avert one breast cancer death, and stage at diagnosis, finding …

  • Annual screening starting at age 40 had the biggest mortality reduction compared to no screening, at 7.9 fewer deaths per 1,000 women, compared to biennial 40-74 (5.9) and biennial 50-74 (4.6) 
  • Annual screening from 40-74 had the lowest number of women who must be screened to avert one death (127) compared to biennial 40-74 (169) and biennial 50-74 (220)
  • Earlier annual screening would produce the greatest stage shift to more early invasive (stage 1 and stage 2a) cancers detected compared to other regimens 

The Takeaway

The Canadian task force is expected to complete its review by the end of the year – where it will land on the issue is anyone’s guess. It’s hoped that the new study – as well as other research on mammography’s effectiveness in Canada published in the last couple years – will spur the group to lower the screening age. But breast imaging experts we spoke with are skeptical given the task force’s preference for randomized clinical trials, which haven’t been performed in Canada on breast screening in decades.

Get every issue of The Imaging Wire, delivered right to your inbox.

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!