{"id":19609,"date":"2026-09-28T05:14:44","date_gmt":"2026-09-28T08:14:44","guid":{"rendered":"https:\/\/rtmedical.com.br\/tmp-en-1790583283439\/"},"modified":"2026-09-28T05:14:50","modified_gmt":"2026-09-28T08:14:50","slug":"sustainable-lung-nodule-follow-up","status":"publish","type":"post","link":"https:\/\/rtmedical.com.br\/en\/sustainable-lung-nodule-follow-up\/","title":{"rendered":"Why Lung Nodule Follow-Up Programs Fail to Scale"},"content":{"rendered":"

Lung nodule follow-up<\/strong> programs fail to scale because every finding triggers a chain of manual tasks \u2014 image review, patient communication, interval selection, ordering, tracking \u2014 and no department has enough staff to close all of them. That is the core argument of “Sustainable Screening,” an opinion piece published September 25 in Radiology Today<\/em> by Chris Wood, CEO of RevealDX, a company that sells radiomics software for lung nodules. The thesis deserves attention, but it must be read with the byline in mind: some of the numbers come from the literature, some have no traceable source, and the proposed fix is the product the author sells.<\/p>\n

\"Bar
The report gets it right; the system does not execute. Chain of adherence across two US Veterans Affairs systems. Data: Hedstrom et al., Annals ATS, 2022.<\/figcaption><\/figure>\n

What the op-ed argues<\/h2>\n

Wood starts from a familiar diagnosis: lung cancer remains the leading cause of cancer death in the United States, and low-dose CT screening, although well established, reaches few people. He cites a CDC estimate that only about 18% of eligible adults have been screened. He adds that “much more than half” of those diagnosed do not meet screening criteria, points to women with breast cancer as a candidate group for inclusion, and notes that CT scans ordered for other reasons \u2014 abdominal pain, cardiac workup, infection \u2014 find nodules in people who are often at high risk.<\/p>\n

Then comes the heart of the piece. According to the author, up to 40% of chest CTs identify a nodule, only 22% of patients with small nodules follow up on time, and half of lung cancers “upstage” during the interval before the follow-up scan. The conclusion: if each incidental finding requires a proportional increase in manual review and navigation, programs will not grow. The proposed way out is radiomics and AI to separate patients who warrant attention from those who can stay in routine surveillance \u2014 without the tool making the final call, Wood stresses.<\/p>\n

What the literature confirms \u2014 and what it does not<\/h2>\n

We checked each figure. The participation number holds up: the American Cancer Society analysis of the 2024 National Health Interview Survey, published in JAMA<\/em> in November 2025, found 18.7% of eligible adults up to date with screening, out of 12.76 million people eligible under USPSTF criteria. The same study estimated 62,000 avoidable deaths over five years at full coverage. We covered the trend when we reported that CT lung screening uptake is rising<\/a>, albeit from a low base.<\/p>\n

The claim that “much more than half” of cases fall outside the criteria overstates what the studies we found show. A nationally representative US analysis of adults aged 50 and over diagnosed with lung cancer found that 52% met the 2021 USPSTF criteria \u2014 meaning roughly half were excluded. Smaller hospital cohorts report eligible fractions between 35% and 64%, depending on the population. The author’s point holds in direction, not in magnitude.<\/p>\n

The “up to 40%” nodule prevalence is also above the largest published series. The study by Gould and colleagues in the American Journal of Respiratory and Critical Care Medicine<\/em>, covering 415,000 chest CTs at Kaiser Permanente between 2006 and 2012, found nodules on 24% to 31% of exams, with the trend rising. In low-dose screening, European programs such as NELSON reported higher fractions, which may explain the 40% ceiling.<\/p>\n

The two most striking numbers \u2014 22% timely follow-up and 50% upstaging \u2014 we could not locate in a primary source, and the piece does not reference them. The literature shows something similar but measured differently. Hedstrom and colleagues, writing in the Annals of the American Thoracic Society<\/em> in 2022 on 4,586 patients across two Veterans Affairs systems that already had nodule tracking, found that radiologist reports followed Fleischner guidelines 86.6% of the time, yet the follow-up actually completed matched the guideline in only 46% of cases and matched the report’s recommendation in 54.3%. In nearly 40% of cases the follow-up was less aggressive than indicated. On stage, the SUMMIT study published in CHEST<\/em> in 2023 found that delays of three to six months in the interval scan did not change stage at diagnosis, while delays beyond six months roughly tripled the relative risk of stage II or higher disease \u2014 with a small sample and borderline significance. In Denmark, Borg and colleagues reviewed 4,066 stage IV cases and concluded that 45% of previously described nodules had not been appropriately monitored.<\/p>\n

Fleischner, Lung-RADS and the gap between them<\/h2>\n

Part of the complexity Wood describes stems from having two management systems. For incidental nodules in adults aged 35 and over, the 2017 Fleischner Society guidelines apply: a single solid nodule under 6 mm needs no routine follow-up in a low-risk patient; 6 to 8 mm calls for CT at 6 to 12 months; above 8 mm, CT at 3 months, PET\/CT or tissue sampling. For nodules found in a screening program, Lung-RADS applies, with categories ranging from annual continuation to immediate workup. Intervals differ, thresholds differ, and a 7 mm nodule can be managed differently depending on which door it came through.<\/p>\n

This matters because an incidental finding belongs to no program. In screening there is a coordinator, a registry and a return schedule. In the incidental setting there is a sentence at the end of a report that the ordering clinician \u2014 often a cardiologist or surgeon who requested the scan for another reason \u2014 may not read as a priority. We showed the extreme cost of that failure in the case of the 8 mm nodule that became stage 4 cancer<\/a> nearly three years after being correctly described.<\/p>\n

Add the variability of the report itself. A study we covered in August showed that radiologists vary widely when reading the same lung CT<\/a>, including in how they phrase the recommendation. A free-text recommendation with no code and no target date is invisible to any automated system.<\/p>\n

Where AI and radiomics actually fit<\/h2>\n

Here it helps to separate two things the op-ed blends: tracking open items and stratifying risk. The first is a workflow problem. Tools that extract recommendations from report text, build a worklist in the RIS or PACS and escalate overdue items to a patient navigator already exist and have a clear return. They require no radiomics; they require a structured field and an owner for the queue.<\/p>\n

The second is the author’s company’s territory. Software that computes malignancy probability from nodule texture on CT is a regulatory reality: Optellum’s Virtual Nodule Clinic received FDA 510(k) clearance in March 2021 with a 1-to-10 score; RevealDX’s RevealAI-Lung was cleared in February 2026 and produces a malignancy similarity index. The company states, in its own materials, that the index brought forward 40% of diagnoses and cut false positives by 22% in validation cohorts \u2014 vendor figures, not independent trial results. It is also worth recalling that AI in screening still performs unevenly in real-world use, as we discussed when covering how cancer survival rises as AI enters screening<\/a>.<\/p>\n

A well-calibrated risk score can cut unnecessary follow-up for low-probability nodules and bring forward biopsy for high-probability ones. But it does not solve the 46% concordance bottleneck: a patient whose follow-up was never scheduled remains unscheduled, score or no score. The logical order is to close the loop first and stratify second.<\/p>\n

What changes for Brazil and Latin America<\/h2>\n

Brazil has no national lung cancer screening program. Bill 2550\/2024 is pending in Congress, and the 2024 joint recommendation by the Brazilian thoracic surgery, pulmonology and radiology societies proposes low-dose CT for people aged 50 to 80 with a significant smoking history, but without systematic funding. ProPulm\u00e3o, run by Hospital Israelita Albert Einstein, is the longest-running initiative; its mobile arm triaged 5,223 people in Bahia between 2023 and 2025 and performed 2,018 CT scans. We discussed strategies for expanding coverage in our WCLC 2026 coverage<\/a>.<\/p>\n

In practice, the vast majority of nodules in Brazil and across the region are incidental, found on CTs ordered for another reason, and land in a system that rarely connects report, requester and schedule. In the public system, authorization for a follow-up scan can take months \u2014 which, on the SUMMIT data, is exactly the window in which stage starts to shift. Before any radiomics software, the highest-impact step for a Latin American department is modest: standardize the recommendation phrasing, tag the report with a retrievable code and check monthly how many recommendations became exams.<\/p>\n

How to read an op-ed written by the vendor<\/h2>\n

None of this invalidates Wood’s thesis. Nodule programs genuinely do not scale by adding people, and image-based risk stratification is one of the few levers that reduces work rather than adding it. But the most striking numbers in the piece are unsourced, and the proposed remedy coincides with the product of the company the author runs \u2014 a product that gained Medicare reimbursement under specific CPT codes after FDA clearance. Readers should take the structural argument, check the data against the literature, and demand independent validation before buying the tool.<\/p>\n

Source:<\/strong> Radiology Today<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"

RevealDX’s CEO argues AI makes nodule follow-up sustainable. We checked the numbers: 18.7% screened, 46% Fleischner concordance, delays upstage cancer.<\/p>\n","protected":false},"author":1,"featured_media":19583,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"om_disable_all_campaigns":false,"_monsterinsights_skip_tracking":false,"_monsterinsights_sitenote_active":false,"_monsterinsights_sitenote_note":"","_monsterinsights_sitenote_category":0,"ngg_post_thumbnail":0,"_rt_cluster":"","fifu_image_url":"","fifu_image_alt":"","footnotes":""},"categories":[102,100],"tags":[],"class_list":["post-19609","post","type-post","status-publish","format-standard","has-post-thumbnail","category-ai","category-radiology"],"aioseo_notices":[],"rt_seo":{"title":"","description":"RevealDX's CEO argues AI makes nodule follow-up sustainable. We checked the numbers: 18.7% screened, 46% Fleischner concordance, delays upstage cancer.","canonical":"","og_image":"","robots":"index,follow","schema_type":"Article","include_in_llms":true,"llms_label":"Sustainable lung nodule follow-up: RevealDX op-ed fact-checked","llms_summary":"Critical read of the 'Sustainable Screening' op-ed (Chris Wood, CEO of RevealDX, Radiology Today, Sep 2026) arguing radiomics and AI make lung nodule programs scalable. Confirmed: 18.7% of eligible US adults screened (ACS\/NHIS 2024, JAMA 2025); Fleischner adherence falls from 86.6% at the report to 46% at completed follow-up (Hedstrom, Annals ATS 2022); delays >6 months triple the risk of stage \u2265II (SUMMIT, CHEST 2023). Not found in primary sources: 22% timely follow-up and 50% upstaging. Context on Fleischner vs Lung-RADS, RevealAI-Lung (FDA Feb 2026), Optellum, and Brazil's lack of a national program.","faq_items":[],"video":[],"gtin":"","mpn":"","brand":"","aggregate_rating":[]},"_links":{"self":[{"href":"https:\/\/rtmedical.com.br\/en\/wp-json\/wp\/v2\/posts\/19609\/"}],"collection":[{"href":"https:\/\/rtmedical.com.br\/en\/wp-json\/wp\/v2\/posts\/"}],"about":[{"href":"https:\/\/rtmedical.com.br\/en\/wp-json\/wp\/v2\/types\/post\/"}],"author":[{"embeddable":true,"href":"https:\/\/rtmedical.com.br\/en\/wp-json\/wp\/v2\/users\/1\/"}],"replies":[{"embeddable":true,"href":"https:\/\/rtmedical.com.br\/en\/wp-json\/wp\/v2\/comments\/?post=19609"}],"version-history":[{"count":1,"href":"https:\/\/rtmedical.com.br\/en\/wp-json\/wp\/v2\/posts\/19609\/revisions\/"}],"predecessor-version":[{"id":19613,"href":"https:\/\/rtmedical.com.br\/en\/wp-json\/wp\/v2\/posts\/19609\/revisions\/19613\/"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/rtmedical.com.br\/en\/wp-json\/wp\/v2\/media\/19583\/"}],"wp:attachment":[{"href":"https:\/\/rtmedical.com.br\/en\/wp-json\/wp\/v2\/media\/?parent=19609"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/rtmedical.com.br\/en\/wp-json\/wp\/v2\/categories\/?post=19609"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/rtmedical.com.br\/en\/wp-json\/wp\/v2\/tags\/?post=19609"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}