AI radiology insights
Clinical evidence, compliance guides, and operational benchmarks for radiology practices navigating AI CT reporting
A Hospital Bought a Radiology Practice. Wait Times Got Worse.
Intermountain Health closed its acquisition of Las Vegas's Steinberg Diagnostic in January. By August, some patients were waiting three weeks for scan results the practice used to turn around in under ten days. What the case shows about radiology practice consolidation and capacity.
Read article →The $2.6M Radiology Billing Gap Is Also a Reporting Problem
A new analysis puts the annual revenue-cycle opportunity at a mid-size radiology practice at $2.6 million, most of it from denials and underpayments. The industry's own denial data points at documentation — not just coding — as a recurring cause.
Read article →More Than 1 in 5 Neuroradiology Second Opinions Turn Up a Major Discrepancy
A 580-case UK study found subspecialist neuroradiology second opinions changed the read in 42% of cases — and 21% involved a major, management-changing discrepancy. What that means for how often patients actually get a real second read.
Read article →ECRI's New AI Error Tracker Changes the Radiology AI Vendor Evaluation Checklist
ECRI now tracks AI errors and near-misses across every AI vendor. Here's what belongs on your radiology AI vendor evaluation checklist now.
Read article →A $2.54M Glioblastoma Imaging Grant Points to the Real Bottleneck: Reporting Extent, Not Just Detection
Wayne State researchers won a $2.54M NIH grant to build a quantitative PET tracer for glioblastoma. The real story isn't detection — it's why turning a scan into a number surgeons can act on is the hard, underfunded part of neuro-oncology imaging.
Read article →Incidental Renal Mass on CT: Why the Report Matters
The FDA just fast-tracked a PET agent for indeterminate renal masses. But most incidental kidney findings never reach any second test — they live or die on how precisely the original CT or MRI report describes them.
Read article →42% of Millennial Radiologists Say Pay Stagnated: What the Data Shows
A new Medscape survey finds 42% of millennial radiologists saw pay stagnate or fall in 2025 — evidence that throughput, not headcount, is the lever left.
Read article →QPA Formula Struck Down: What It Means for Radiology
An appeals court vacated the QPA formula insurers use under the No Surprises Act. What changes for radiology out-of-network reimbursement leverage.
Read article →Cloud-Based Medical Image Sharing: The Next Gap
Epic's new one-click tool moves full-resolution scans between hospitals without CDs. The radiology report riding along still has no standard structure.
Read article →The Coalition for Health AI's New Security Work Group: A Vendor Checklist for Imaging Centers
CHAI formed a work group to build playbooks against frontier-model cyber risk in clinical AI. What it means, and the questions imaging centers should ask an AI-reporting vendor before signing.
Read article →St. Luke's $23M Viewer Deal Shows Where Radiology IT Money Still Isn't Going
St. Luke's $23M Pro Medicus deal buys a cloud imaging viewer, not reporting AI. What the purchase says about where radiology IT budgets go — and don't.
Read article →AI Malignancy Risk Models for Incidental Lung Nodules
A multicentre European Radiology study tested two deep learning models against the Brock risk score on 269 incidental lung nodules from three hospitals. The AI models beat Brock and stayed consistent site to site.
Read article →Radiology AI and Clinical Outcomes: The Evidence Gap
A radiologist survey finds AI adoption in breast imaging outpaces measurable outcomes: fewer than half report a real benefit on recall rates or burnout.
Read article →LLMs Beat Clinicians at Writing Radiology Order Indications
A 77,626-exam UCSF study found LLMs write more comprehensive, more factual clinical indications than clinicians. Why that gap matters for AI report accuracy.
Read article →Why the Words Radiologists Use Can Delay Care: What New Research Shows
A new CT study found imprecise wording drove indeterminate reports, and vague reports led to surgery more often. What it means for radiology report language.
Read article →Can an LLM Catch Radiology QC Errors? New Study
A 735-patient study found an LLM beat manual reviewers on accuracy, finishing 50 reports in 13 vs 213 minutes — what it means for radiology quality assurance.
Read article →Incidental Findings on Chest CT: The Breast Lesions Radiologists Are Missing
A new Academic Radiology study of ED chest CTs found that 70% of scans with a visible, later biopsy-confirmed breast cancer went unreported at the initial read. Here's what it shows about incidental findings on chest CT.
Read article →How to Choose a Teleradiology Company: A Buyer's Guide
How to choose a teleradiology company: prelim vs final reads, turnaround SLAs, subspecialty coverage, QA, pricing, and the questions to ask before signing.
Read article →Congress Wants to Pay for FDA-Cleared Imaging AI. Here’s Why That Matters Beyond Veterans.
A bipartisan bill would put $25M behind FDA-cleared lung-imaging software for veterans. What direct federal AI-imaging funding signals for CT reporting.
Read article →AI in Radiology: The Complete 2026 Guide
A complete 2026 guide to AI in radiology: detection vs. drafting vs. triage, the FDA/CE regulatory landscape, the evidence base, and how practices adopt it safely.
Read article →Nighthawk Radiology: History, Economics, and the AI Shift in After-Hours Coverage
Nighthawk radiology explained: the history of NightHawk Radiology Services, how overnight preliminary reads work, the economics of after-hours coverage, and how AI drafting changes the math.
Read article →Teleradiology Jobs: A 2026 Careers Guide for Radiologists
Teleradiology jobs, explained for radiologists: how remote reading works, pay versus on-site roles, multi-state licensing and the Interstate Medical Licensure Compact, and what groups look for in 2026.
Read article →The Patients First Act: What It Means for Radiology Pay
The bipartisan Patients First Act would link Medicare physician pay to inflation, cap conversion-factor cuts at 2.5%, and revive imaging appropriate use criteria through the ROOT Act. What the legislative fix means for independent radiology groups.
Read article →The $500M Imaging Deal and the Capital Gap Smaller Providers Face
GE HealthCare and Catholic Health signed a 10-year, $500M imaging Care Alliance across 40+ sites. What enterprise imaging modernization means for smaller providers — and how asset-light AI CT reporting narrows the capital gap.
Read article →Automation Bias in Radiology: The Case for Human Review
A 2026 RSNA Radiology eye-tracking study found radiologists' sensitivity fell from 71% to 39% when AI missed a cancer. Here's what automation bias means for AI reporting — and how to evaluate a vendor's safeguards.
Read article →Radiology Reporting: A Complete Guide
Radiology reporting explained: narrative vs structured reports, ACR and RSNA standards, turnaround, and how AI drafting is changing the report workflow.
Read article →What Is Teleradiology? How It Works in 2026
Teleradiology explained: how remote radiology reading works in 2026 — preliminary vs final reads, nighthawk history, licensing, economics, and where AI fits.
Read article →2027 Medicare Physician Fee Schedule: What It Means for Radiology
The proposed 2027 Medicare Physician Fee Schedule cuts the conversion factor 1.68%, yet CMS estimates a net +2% overall impact for radiology. What the numbers mean for imaging-group margins — and why revenue per study, not headcount, is the real story.
Read article →Will AI Replace Radiologists? An Honest, Data-Led Answer
Will AI replace radiologists? The data says no. What AI does today (drafting, triage), what it can't (accountability, the signature), and what the workforce numbers actually show.
Read article →CT Scans for Pulmonary Embolism: When Cancer Patients Can Safely Skip Them
A JAMA randomized trial found the YEARS algorithm safely avoided a CT scan for pulmonary embolism in 22% of cancer patients. Here's what it means for CT volume and reporting.
Read article →Low-Value Imaging: What Clinician Knowledge Reveals About Appropriate Use
A 2026 JAMA Internal Medicine study of nearly 900,000 Medicare beneficiaries found that patients of physicians in the top quartile of clinical-knowledge scores were measurably less likely to receive low-value imaging. Why appropriate use — not raw volume — is the demand-side quality problem, and where structured AI CT reporting fits.
Read article →Radiology and Private Equity: How Independent Groups Can Stay Independent
A network of oncologists just banded together to avoid selling to private equity — and radiology faces the same squeeze. PE firms acquired 151 U.S. radiology practices between 2013 and 2023. Beyond structural pacts, there's an operational lever independent groups can pull: adding read volume and after-hours coverage with AI CT reporting, without surrendering equity.
Read article →Fewer Imaging Gatekeepers, More Scans: The Capacity Squeeze
A viral clinician thread asked whether the imaging-cautious physician is a "dying breed." Behind the anecdote is a structural problem: ED CT use per Medicare beneficiary rose 95.8% in a decade while ED visits fell 16%. What overutilization of medical imaging means for radiologist capacity — and how AI CT reporting absorbs the overflow without proportionally growing headcount.
Read article →Only 48% of Radiologist Job Listings Show Pay — What That Signals
Only about 48% of U.S. radiologist job listings include a salary estimate, a July 2026 analysis of 5,000+ postings found. What opaque pay reveals about a supply-constrained market — and why imaging centers should treat AI CT reporting as capacity relief, not a hiring race.
Read article →The Best Metro Areas for Radiologists in 2026 — and the Access Gap Behind the Rankings
A new Marit Health analysis ranks Minneapolis-St. Paul the top U.S. metro for radiologists, ahead of Dallas-Fort Worth and Portland. Here's what the geography of radiologist supply means for community and rural imaging centers that can't win the metro talent war.
Read article →Teleradiology Just Got Its Own Lobby: A Policy Watch-List for Teleradiology Companies
RADPAC — America's largest radiology PAC — just launched a subcommittee focused entirely on teleradiology advocacy. The issue list surfacing around its launch (licensure compacts, CMS supervision, offshore reading, AI accountability) is the closest thing yet to a policy radar for teleradiology companies and the groups that depend on remote reads.
Read article →Integrating Breast and Lung Cancer Screening: The Operational Playbook
A new JACR analysis from Thomas Jefferson University finds only 54% of women in a lung cancer screening program had a screening mammogram within a year of their low-dose CT — versus a national estimate of almost 80% within two years. The authors call for integrated, one-stop screening. Here's what combined breast and lung screening asks of an imaging center's scheduling, eligibility capture, and reporting capacity.
Read article →When Radiology Outsourcing Goes Wrong: Anatomy of a Failed Teleradiology Transition
A Tennessee health system replaced its local radiology group with an overseas teleradiology company — within days, STAT scans waited up to six hours and non-radiologists were doing preliminary reads. A failure-mode analysis, a due-diligence checklist for any outsourcing contract, and the alternative that keeps turnaround control in-house.
Read article →Site-Neutral Payments, Explained: What CMS's Proposed $260M Imaging Cut Changes
CMS's proposed 2027 OPPS rule would pay grandfathered off-campus hospital departments physician-office rates for imaging without contrast — about 40% of the current hospital rate, a $260 million first-year cut. What site-neutral payments are, who wins and who loses, and why per-study reporting cost becomes the margin lever both sides can control.
Read article →A Server Glitch Made Radiologists Read the Wrong Patient. What It Means for AI Reporting Pipelines
An FDA Class 2 recall of 340 GE HealthCare AW Server units shows how a silent worklist defect can open the previous patient's exam with no warning. Why AI CT reporting pipelines need hard patient-context integrity checks and a mandatory sign-off gate.
Read article →Anatomy of a $7M Missed-Cancer Verdict — and Where AI Reporting Fits in the Liability Picture
A Florida jury awarded nearly $7M after a palpable breast lump reported as benign turned out to be terminal cancer. A neutral look at the case — and where AI CT reporting sits in the malpractice picture: a second-read safety net, not autonomous diagnosis.
Read article →Simpler Lung Cancer Screening Criteria Could Mean a Lot More Chest CTs
A new JAMA Internal Medicine study finds a simple 'years smoked' threshold captures 97% of the highest-benefit patients versus 77% under current USPSTF pack-year criteria — and could roughly double the eligible population. Here's what broader eligibility means for low-dose chest CT volume and reporting capacity.
Read article →Medical Device Cybersecurity: What the CISA DICOM Advisory Means for AI Imaging Buyers
CISA's June 2026 advisory flagged five vulnerabilities in OFFIS DCMTK, an open-source DICOM toolkit embedded across imaging software. Here's what it means for imaging IT — and the security questions to ask any AI CT reporting vendor about data handling, PHI flow, and deployment model.
Read article →Radiology Prior Authorization Reform: What Faster Medicare Advantage Approvals Mean for Imaging Throughput
A House committee advanced the Improving Seniors' Timely Access to Care Act (H.R. 3514) to curb prior authorization in Medicare Advantage. Faster approvals mean more scans reach the reading room — moving the bottleneck downstream to reporting turnaround.
Read article →Should Patients Be Told When AI Reads Their Scan? What a New Survey Reveals
In a survey of more than 1,000 imaging patients, 96% said they should be told when AI is used to report on their scan — and 64% said both the doctor and the AI share the blame if it's wrong. Here's what the data means for how imaging centers disclose AI and keep a radiologist accountable.
Read article →AI Radiology Reporting: What Chest X-ray Studies Show About Draft-Then-Sign
Generative AI report drafting has arrived. But two peer-reviewed chest X-ray studies show the model works as a first-draft engine a radiologist reviews and signs — cutting reading time and lifting sensitivity — not as an autonomous reader. Here's what that means for AI radiology reporting and quality improvement.
Read article →Who Gets Radiology AI? Why Reimbursement Design Could Deepen Healthcare Disparities
New Neiman Institute research finds Medicare's add-on payment for stroke AI reached just 21% of eligible cases at its 2022 peak — concentrated at large stroke centers, while hospitals in more deprived areas were less likely to use it. Here's how reimbursement design gates access to radiology AI, and what a non-capital, per-study model changes for smaller providers.
Read article →AI Cut a 37-Hospital System’s MRI Wait Times by More Than 60% — But Faster Scans Just Move the Bottleneck
A 37-hospital system halved MRI scheduling delays with FDA-cleared acquisition-speed AI. But faster acquisition pushes the constraint downstream to reporting. Here’s the throughput case for pairing acquisition AI with AI report drafting to clear the backlog end-to-end.
Read article →Foundation Models vs Narrow AI in Radiology: Why One Model Beats 30 Detection Tools
Buy narrow AI and you end up with seven detection tools on a single chest CT — and still no report. Foundation models flip the architecture: one system, one complete ready-to-sign draft. Here is the published evidence behind the shift, what it does to cost and radiologist workload, and the three questions to ask before you buy.
Read article →Should Radiology AI Be Priced on Results? The Case for Performance-Based Pricing
Today you pay per study whether the AI helps or not — and when it is wrong, you absorb the cost of re-reading and fixing it. Here is the case for tying radiology AI pricing to performance, what a threshold-based model could look like, and why it all comes down to trust.
Read article →AI Radiology Reporting Software: A 2026 Buyer's Guide for Imaging Centers
Not all AI radiology tools produce the same output. Some flag findings; others deliver complete signed reports. Here's how to evaluate vendors — accuracy data, pricing models, radiologist review, quality guarantees, and compliance — before you commit.
Read article →How to Switch from Teleradiology to AI CT Reporting: A Step-by-Step Guide
DICOM integration completes in under one week. A structured pilot lets you validate report quality before any contract change. Here is the complete transition process — from pilot evaluation to full cutover — including what your IT team actually needs to do.
Read article →AI Radiology for Small and Community Hospitals 2026: Coverage Options, Costs, and Implementation
Small hospitals and critical access hospitals face the same problem: can't hire a full-time radiologist, can't afford locum rates, and traditional teleradiology after-hours surcharges make 24/7 coverage unaffordable. Full comparison of coverage options, costs, and implementation path in 2026.
Read article →AI Radiology Terminology Glossary: Key Terms Explained
Reference guide to 18 key terms in AI CT reporting, teleradiology, and medical imaging — from DICOM and HL7 to sensitivity/specificity, foundation models, and after-hours surcharges. Plain-language definitions with clinical and operational context.
Read article →After-Hours Radiology Coverage Options 2026: On-Call, Locum, Teleradiology, and AI Compared
Traditional teleradiology charges 30–100% surcharges for after-hours CT reads. A center reading 500 after-hours studies per month can pay $90,000–$300,000 per year in surcharges alone. Full comparison: in-house on-call, locum, traditional teleradiology, and AI CT reporting — costs, availability, quality, and guarantees.
Read article →CT Radiology Coverage Costs 2026: In-House, Teleradiology, Locum, and AI Compared
A mid-volume outpatient imaging center can spend $300,000–$1.2 million annually on CT radiology coverage — depending entirely on the model. Full cost breakdown: in-house radiologist, locum, traditional teleradiology, and AI CT reporting, with per-study rates, after-hours costs, and quality guarantees.
Read article →How Accurate Is AI Radiology Reporting? Evidence from Published Clinical Studies
What does the peer-reviewed evidence say about AI CT reporting accuracy? We analyzed two independent clinical studies — including a retrospective evaluation of 90 emergency chest CT scans — and compared the numbers to traditional radiology benchmarks.
Read article →Radiologist Shortage 2026: How AI CT Reporting Fills the Gap
The US faces a projected a shortage of up to 86,000 physicians by 2036, with radiology among the hardest-hit specialties. Here's what the data says — and how outpatient centers and teleradiology providers are using AI to cover the gap today.
Read article →CT Report Turnaround Time Benchmarks 2026: What's Normal and What's Not
ACR guidelines say routine CT reads should be signed within 24 hours. The reality is often 36–72 hours. Here's what drives turnaround times, what benchmarks look like across practice types, and what AI-assisted reporting actually delivers.
Read article →AI Teleradiology vs Traditional Teleradiology: Full 2026 Comparison Guide
Traditional teleradiology services charge $80–350 per study and take 4–24 hours. AI-assisted teleradiology delivers the same output in 2–12 hours at lower per-study cost. But there are real differences worth understanding before you switch.
Read article →Is AI Radiology Reporting HIPAA Compliant? What to Ask Before You Buy
HIPAA compliance for AI radiology goes beyond encrypting images. A Business Associate Agreement, US-based infrastructure, audit logs, and radiologist sign-off are all required. Here's the compliance checklist — and what xAID satisfies.
Read article →