The scale of the donation reflects an unusual fundraising model in which corporate giving amplifies grassroots employee participation. This year’s campaign included more than $766,000 in contributions from employees of the Energy Transfer family of partnerships, which the company matched with corporate funds at a rate of 50 percent. Since the effort began in 2014, employees have led fundraising initiatives that have resulted in more than $20 million donated to UT MD Anderson, and the combined total of company support now exceeds $21 million. That longevity matters in biomedical research, where multi-year funding streams allow laboratories to plan experiments, retain specialized staff, and pursue lines of investigation that may take years to bear fruit.

Leadership at Energy Transfer framed the gift as an expression of a shared institutional mission rather than a one-time gesture. Tom Long, co-Chief Executive Officer of Energy Transfer, said the commitment employees have shown to UT MD Anderson over the years reflects a shared belief in the power of innovation to change lives. He noted that the breakthroughs happening at the cancer center, and the hope they bring to patients and their families, reinforce the importance of the collaboration. Such statements underscore a growing trend in which private companies channel philanthropic resources toward translational biomedical science, bridging gaps that conventional grant funding does not always cover.

At MD Anderson, the funds arrive at a moment when pediatric oncology is undergoing a significant transformation. Childhood cancers are, in many respects, distinct diseases from their adult counterparts, driven by different genetic alterations and arising in tissues that are still developing. That biological reality means that treatments designed and tested primarily in adults cannot simply be transplanted into pediatric care without careful adaptation. Researchers in pediatric divisions therefore focus on identifying the molecular vulnerabilities unique to childhood tumors, developing age-appropriate dosing and treatment protocols, and minimizing the long-term side effects that can follow young patients for decades after they are declared cancer free.

Richard Gorlick, M.D., division head and chair of Pediatrics at UT MD Anderson, emphasized the role that sustained private generosity plays in that enterprise. He said that Energy Transfer, Sunoco and USA Compression employees continue to demonstrate an extraordinary commitment to children and families facing cancer, and that their generosity helps researchers pursue new discoveries and translate those advances into better treatments and outcomes for patients across the country. The language of translation is significant: moving a laboratory finding, such as the identification of a drug target or a new biomarker, into a clinically validated therapy requires resources that sit outside the standard cycle of federal research grants, particularly during the early and risky stages of development.

The human dimension of that research pipeline was on display in this year’s campaign through its patient ambassador, Braeden Cochran, whose medical journey illustrates how clinical research can change the trajectory of a young life. Cochran was diagnosed as a college freshman with a rare sub-type of blood cancer known as non-Hodgkin B-cell lymphoma, a disease in which immune cells called B cells grow uncontrollably. He began treatment at UT MD Anderson in 2024, a point he describes as a turning point in his care. Now 21, he has become a public face of the fundraising effort, demonstrating to employee donors the concrete consequences of their contributions.

Cochran’s treatment included participation in an immunotherapy-based clinical trial, an approach that has reshaped the treatment landscape for several blood cancers in recent years. Immunotherapy works differently from conventional chemotherapy: rather than directly poisoning rapidly dividing cells, it recruits or engineers the patient’s own immune system to recognize and attack malignant cells. For patients, this can mean the ability to tolerate comprehensive treatment protocols more effectively, with different and often more manageable side effects than those associated with intensive traditional regimens. Cochran credited the trial with enabling him to better tolerate his full course of treatment, an outcome that reflects the careful design work that goes into matching patients with experimental approaches suited to their disease.

Equally telling is what Cochran said about the environment in which he received care. He recalled that he did not really have the same confidence in his own healing until he came to UT MD Anderson, and that was where things really started to turn around for him. He described the experience of standing in his hospital room and seeing the research buildings on the same campus just outside, noting that it was encouraging to know that work on new treatments was being done all the time. That physical proximity between bedside and laboratory is a defining feature of comprehensive cancer centers, where clinicians and researchers operate in parallel and patient needs can inform research priorities in near real time.

Today, Cochran is cancer free and back in school, an outcome he attributes to the advanced lifesaving treatment options and expert care he received at the institution. Stories of this kind carry weight beyond their emotional resonance. For the thousands of employees who participate in the annual giving campaign, a named ambassador provides a tangible connection between payroll deductions and scientific progress, and for the researchers who depend on philanthropic support, patient outcomes serve as the clearest possible justification for continued investment in early-stage and high-risk science.

The $2.2 million gift thus arrives as both a financial milestone and a case study in how modern cancer research is funded. Federal agencies remain the largest source of basic biomedical research support, but philanthropy increasingly fills critical roles: seeding pilot studies, supporting clinical trials that test novel immunotherapy approaches, funding young investigators, and sustaining the infrastructure that allows discoveries to move from bench to bedside. With more than $21 million contributed since 2014 and this year’s record total, the Energy Transfer family of partnerships has built one of the more enduring corporate-employee philanthropic partnerships in pediatric oncology. For the children and young adults treated at UT MD Anderson, and for the researchers racing to develop the next generation of therapies, the donation represents something more than a number: it is continued momentum for a scientific enterprise whose ultimate measure is measured in patients who, like Cochran, return to their studies and their lives after cancer.

Subject of Research: Philanthropic funding of pediatric cancer research and immunotherapy clinical trials at UT MD Anderson Cancer Center

Article Title: Energy Transfer and Sunoco donate $2.2 million to support pediatric cancer research at UT MD Anderson

Article References: Energy Transfer and Sunoco donate $2.2 million to support pediatric cancer research at UT MD Anderson. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: pediatric cancer, MD Anderson, Energy Transfer, Sunoco, philanthropy, immunotherapy, clinical trial, non-Hodgkin lymphoma, employee giving campaign, cancer research funding, pediatric oncology, translational research

Cachexia is a severe and progressive wasting syndrome characterized by unintentional weight loss and the erosion of skeletal muscle, and it affects a large proportion of patients with advanced cancers. Unlike ordinary weight loss, cachexia cannot be reversed simply by eating more or by nutritional supplementation alone. The syndrome involves deep metabolic disruption: inflammation, altered metabolism in the liver, and signals traveling between the brain and the body’s organs conspire to break down muscle tissue even when patients consume adequate calories. In patients with metastatic pancreatic cancer, cachexia is especially common and can become the factor that limits how long a person is able to continue chemotherapy, the treatment that offers the best chance of extending life.

The scale of the problem is sobering. Up to half of all cancer deaths are attributable to cachexia, making it a silent contributor to mortality that often goes unaddressed in treatment planning. Patients describe profound fatigue, loss of strength, and a body that seems to be consuming itself, while families watch loved ones grow frail in ways that chemotherapy alone cannot explain. Because no approved therapies exist, clinicians have been left with supportive care measures that ease symptoms but do not touch the underlying biology. This gap represents what researchers at the Terasaki Institute describe as a major unmet need, and it is precisely the gap that the OnVagus device is designed to address.

The technology takes an unusual approach. Rather than targeting tumors or immune cells directly, the OnVagus device delivers noninvasive electrical stimulation through the skin to modulate the vagus nerve, one of the body’s most important neural highways. The vagus nerve is the longest cranial nerve, connecting the brainstem to organs throughout the chest and abdomen, including the gut and the liver. Through this pathway, the brain constantly monitors and adjusts digestion, metabolism, and inflammatory responses. In cachexia, researchers believe that maladaptive signaling along this axis contributes to the metabolic derangement that drives muscle wasting. By intercepting and modulating those signals, the device aims to intervene in the pathobiology of the syndrome itself rather than merely its symptoms.

The scientific rationale for the approach was established in a preclinical study published last year in the journal Cell, which described both the device concept and its underlying mechanism. That work demonstrated the biological basis for using vagal modulation to influence the wasting process, laying the groundwork for translating the technology into something that could be tested in patients. According to the company, the OnVagus device is the first transcutaneous vagal block device designed for human use, meaning it works entirely through the skin without surgery, implants, or electrodes placed inside the body. This distinguishes it from other neuromodulation approaches that require invasive procedures, a significant consideration for patients whose health is already fragile.

Dr. Aliesha O’Raw, Terasaki Fellow and Co-Founder of OnVagus, emphasized the significance of the noninvasive design. The device, she said, allows the team to modulate the vagus nerve without surgery or implants, providing a new way to intervene in the biological processes driving cachexia. Moving from preclinical work into a first human trial, she noted, is an important step in finding out whether the approach can actually help patients. That transition, from mechanism studies in the laboratory to a device on a patient’s skin, is the kind of translation that medical research institutions aspire to but rarely achieve so directly.

The first step in the clinical program will be a safety lead-in trial enrolling ten pancreatic cancer patients with cachexia at MD Anderson Cancer Center in Houston, one of the world’s leading cancer treatment and research institutions. The first patient is expected to enroll in October. Over eight weeks, the study will assess the device’s safety and usability as its primary outcome, a cautious and appropriate starting point for any first-in-human neuromodulation technology. Secondary measures will include preliminary relief of cachectic symptoms and biomarkers such as liver function, an organ whose metabolic reprogramming is thought to be central to the wasting process. The trial will also collect exploratory measures of progression-free survival and overall survival, offering an early glimpse of whether the intervention might influence disease trajectory and not just body composition.

The trial is supported by a coalition of funders that reflects the multi-stage nature of translational research. Funding comes from the National Cancer Institute’s Small Business Innovation Research program, which supports early-stage medical technology companies; an NIH R01 award to MD Anderson; the American Cancer Society; and BrightEdge, the American Cancer Society’s investment arm. Dr. Xiling Shen, Acting Director of the Terasaki Institute for Biomedical Innovation, described the trial as exactly the kind of translation the Institute wants to see, from a fundamental discovery in the laboratory to a technology being tested in the clinic to help patients. The Institute, a non-profit research organization, focuses on developing biomaterials, cellular and tissue engineering, medical devices, and microfluidic systems, with an explicit mission to accelerate the movement of laboratory discoveries into real-world care.

If the initial safety and usability results are positive, with no serious adverse events observed, the team plans to launch a much larger and more definitive study: a year-long, 120-patient randomized, double-blind trial comparing the active device to a sham device. The sham-controlled design is critical for a neuromodulation therapy, because the placebo effect and the natural variability of cachexia progression could otherwise confound the results. In that pivotal trial, the primary outcome would be a composite measure combining extended time on chemotherapy with reduced cachectic weight loss, an endpoint that captures what matters most to patients, namely the ability to keep receiving effective treatment while preserving their bodies. Secondary measures will include activity levels, mood, and muscle retention, while exploratory measures will track liver function and neurotransmitter levels, potentially illuminating how the intervention works in humans.

For patients with metastatic pancreatic cancer, a disease with notoriously limited treatment options, the prospect of a noninvasive therapy that could preserve strength and quality of life during treatment, and potentially extend survival, is significant. The trial also represents a broader shift in how the medical community thinks about cachexia: not as an inevitable side effect of advanced cancer, but as a treatable condition with its own targetable biology. Much remains to be proven, of course. The safety lead-in is small, and the pivotal trial has not yet begun, so any benefit to patients will only be established through the rigorous clinical testing now getting underway. But the journey of this technology, from a discovery published in Cell to a device cleared for first-in-human testing at one of the world’s premier cancer centers, illustrates how fundamental neuroscience can be redirected toward one of oncology’s most stubborn unsolved problems. If the results hold, the simple act of modulating a nerve through the skin could become a new tool in the fight against the wasting that claims so many lives alongside cancer itself.

Subject of Research: Noninvasive vagus nerve stimulation for cancer cachexia

Article Title: New device targets nerve signals to treat cancer-related wasting, moving to first human trial

Article References: New device targets nerve signals to treat cancer-related wasting, moving to first human trial. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: cachexia, vagus nerve, OnVagus, Terasaki Institute, pancreatic cancer, clinical trial, neuromodulation, MD Anderson, medical device, cancer wasting, NCI SBIR, translational research

The urgency behind the initiative is grounded in stark numbers. Texas currently records a cervical cancer incidence rate of 9.5 per 100,000 women, well above the national figure of 7.5 per 100,000. The state ranks eighth in the nation for incidence and sixth for mortality, a combination that signals not only more diagnoses but also worse outcomes for the women who receive them. The disparities run even deeper at the county level, where some Texas communities report incidence rates between 15 and 21 per 100,000 — as much as five times the elimination threshold. The burden falls disproportionately on uninsured women, rural residents, border communities and minority populations, groups that historically face the greatest barriers to vaccination, screening and follow-up care.

Cervical cancer occupies a unique position in oncology: it is one of the few cancers for which near-complete prevention is scientifically achievable. Virtually every case is caused by persistent infection with high-risk types of human papillomavirus (HPV), a common sexually transmitted virus against which highly effective vaccines have existed since 2006. When precancerous lesions are detected through screening and treated promptly, progression to invasive cancer can be interrupted almost entirely. The World Health Organization has embraced this logic globally, setting the same sub-four-per-100,000 threshold as its international elimination benchmark. The Texas plan translates that global framework into a state-specific strategy, acknowledging that the biology of prevention is settled and the remaining challenge is one of delivery, access and follow-through.

The plan was developed by the Cancer Alliance of Texas Cervical Cancer Elimination Priority Area Workgroup, which grounded its recommendations in a survey of stakeholders across the state and a virtual summit that drew more than 250 participants. The resulting document aligns with multiple priorities from the 2024 Texas Cancer Plan, ensuring that the elimination effort is embedded within the state’s broader cancer control architecture rather than operating as a standalone campaign. More than 40 coalition members have already signed a joint declaration outlining their individual commitments, a structure that distributes responsibility across hospitals, clinics, schools, funders and community organizations instead of concentrating it in a single institution.

At the heart of the roadmap are three pillars, each paired with a measurable target for 2032. The first addresses HPV vaccination: the plan calls for raising the percentage of Texas boys and girls who are up to date on HPV vaccination by age 15 from 50.5 percent to 80 percent. That figure matters because vaccination is most protective when administered well before any potential exposure to the virus, and the age-15 benchmark reflects national guidance emphasizing early completion of the vaccine series. Vaccinating both boys and girls is critical, since herd immunity depends on broad coverage across the population and males can carry and transmit the virus even when they face far lower personal risk of HPV-driven cancers themselves.

The second pillar targets screening, with the goal of increasing the share of Texas women who are up to date on cervical cancer screening from 64 percent to 80 percent. The plan explicitly incorporates two technological advances designed to lower the barriers that keep women from being tested. Primary HPV testing, which uses the virus itself as the screening signal rather than relying solely on cytology, offers improved sensitivity and longer intervals between tests. HPV self-collection, meanwhile, allows women to gather their own samples in a clinical setting or potentially outside traditional exam rooms, an option that has shown promise in reaching women who avoid screening because of discomfort, cost, time constraints or distrust of the medical system. Adding these modalities to the state’s screening toolkit is one of the plan’s most consequential technical commitments.

The third pillar confronts a step that is often overlooked in elimination strategies: what happens after an abnormal result. The plan aims to ensure that 80 percent of women receive the diagnostic follow-up and treatment they need when screening detects something suspicious. This stage is where many prevention programs falter, because an abnormal screening result is only the beginning of a cascade that may involve colposcopy, biopsy and treatment of precancerous lesions. Each additional appointment is an opportunity for patients to be lost to the system, particularly among women juggling work, childcare, transportation challenges or lack of insurance. The plan addresses this vulnerability through community outreach, patient navigation, provider training and survivorship support, pairing clinical capacity with human infrastructure that guides patients through every step.

Patient navigation and access expansion are not confined to a single pillar; they are woven across all three, with particular emphasis on rural and border communities and on including people without health insurance. This design choice reflects the epidemiology of the disease in Texas, where geographic and economic isolation translate directly into higher incidence and mortality. Community health workers, known in many Texas communities as promotores, are positioned as central actors in the strategy, bridging the gap between formal health systems and neighborhoods where medical mistrust, language barriers and logistical hurdles suppress participation in vaccination and screening programs.

Coalition leaders framed the effort in explicitly human terms. Ernest Hawk, M.D., vice president and division head of Cancer Prevention and Population Sciences at UT MD Anderson, said that Texas has the tools to prevent nearly every case of cervical cancer but that tools save lives only when people can utilize them. He described the plan as a shared roadmap to vaccinate more children, screen more women and ensure that every abnormal result leads to care, adding that with innovations like HPV self-collection and the commitment of partners across the state, elimination is within reach. Sharyn Malatok, chair of the Cancer Alliance of Texas, emphasized that elimination will happen in clinics, schools, churches and neighborhoods, through community health workers, promotores and navigators who meet people where they are, and that the plan puts those communities at the center of the effort.

The six-year timeline gives the coalition a concrete horizon: if the 2032 targets are met, Texas would join the small group of jurisdictions that have pushed cervical cancer below the elimination threshold, transforming a disease that once killed thousands of American women annually into a rarity detectable only through surveillance. The full plan is publicly available at TexasCancerPlan.org/CCEP, and its architects are calling on policymakers, health systems, schools, funders and community leaders across the state to join the effort. Whether Texas reaches the finish line will depend on sustained funding, vaccination uptake among the state’s large and diverse pediatric population, and the unglamorous work of making sure no abnormal result slips through the cracks — but for the first time, the state has a single, quantified roadmap for getting there.

Subject of Research: A statewide public health plan to eliminate cervical cancer in Texas through HPV vaccination, screening and timely follow-up care

Article Title: New statewide plan aims to eliminate cervical cancer in Texas by 2032

Article References: New statewide plan aims to eliminate cervical cancer in Texas by 2032. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: cervical cancer, HPV vaccination, cancer screening, Texas, MD Anderson, CPRIT, public health, health disparities, HPV self-collection, patient navigation, cancer elimination, preventive oncology

The technique, known as spinal laser interstitial thermal therapy, or sLITT, involves threading thin laser fibers through a small percutaneous channel directly into a tumor, then destroying it from the inside with precisely controlled heat. Real-time magnetic resonance thermography tracks temperature continuously during the procedure, and an Arrhenius damage model estimates tissue destruction on the treatment workstation as a color-coded overlay on anatomical images. Up to six temperature-monitoring points are placed at the interface between the epidural tumor and the dura mater, the protective membrane covering the spinal cord. When the interface approaches a safety threshold of approximately 50 degrees Celsius, the system automatically shuts the laser off, protecting the cord and nerve roots from thermal injury.

One of the most striking technical challenges the team had to solve involves breathing itself. Respiratory motion degrades the reliability of the thermal map, so ablations are delivered in cycles under ventilatory arrest. Patients are preoxygenated to 100 percent oxygen saturation, each ablation cycle is capped at 100 seconds, and ventilation resumes between cycles. If oxygen saturation falls below 94 percent or the dura-interface temperature threshold is reached, the procedure halts immediately. Each laser fiber produces a zone of thermal damage extending roughly 5 millimeters radially, and additional access cannulas are placed in tandem when greater coverage along the spine is needed. It is a delicate choreography of neurosurgery, physics, and anesthesiology performed inside an intraoperative MRI suite.

The new report, led by Sumeeth V. Jonathan and senior author Claudio E. Tatsui, describes the largest clinical experience with sLITT for chordoma to date: four patients treated across five procedures between January 2016 and November 2025. The tumors spanned the full anatomic range of the disease, including metastatic foci from primary sacral, clival, and cervical chordomas, plus one primary thoracic lesion. Four procedures served as salvage therapy after prior surgery, radiotherapy, or systemic treatment had failed, while one patient received sLITT as primary treatment specifically to avoid the morbidity of open surgery. Median time from diagnosis to LITT was 19.1 months, and all patients entered treatment with excellent functional status, with Karnofsky Performance Status scores of 90 to 100.

Procedurally, the team used a median of four laser fibers and 16 ablations per case, delivering power between 12 and 15 watts. Median case duration was 352.5 minutes, and patients went home after a median hospital stay of just two days. Postoperative stereotactic radiosurgery, a highly focused form of radiation, was administered after all but one procedure, at a median of 7.5 days after LITT. Only one complication occurred: a transient episode of proximal arm radiculopathy that resolved completely with a short course of steroids by the one-month follow-up. There was no procedure-related mortality and no other perioperative complications.

The oncologic results were encouraging. Every patient showed volumetric tumor shrinkage on the earliest postoperative MRI, obtained at a median of 4.2 months, with a mean volume reduction of 25.7 percent. More importantly, by the conventional endpoint of freedom from local progression, meaning no radiographic growth and no further intervention at the treated site, no treated lesion progressed or required additional treatment during available follow-up, which extended to a median of 16.8 months and reached 48.1 months in the longest-followed case. Median overall survival from LITT was 23.0 months. Two patients had died by last follow-up, typically from disease progressing elsewhere in the body, while two remained alive.

A systematic review of the literature performed by the authors found only four previously published case reports of LITT for chordoma, three of which described complete thermal ablation with radiographic response. Those reports used remarkably varied approaches, including endoscopic endonasal, percutaneous translaminar, and transcranial catheter placement, reflecting the anatomic diversity of chordoma across the skull base and spine. No major complications were reported in any of them. The new series more than doubles the aggregate published experience and is the first to report volumetric outcomes and local control systematically across multiple patients.

Two biological questions have long shadowed the idea of laser-ablating chordomas. First, the tumor’s characteristic lipid-rich physaliferous cells were hypothesized to conduct heat poorly compared with other ablated tumor types, potentially limiting ablation size. Yet intraoperative thermography in prior work demonstrated spherical ablation zones of 14 to 16 millimeters in axial diameter, comparable to those achieved in other histologies, suggesting the theoretical concern is not clinically significant. Second, because chordomas are notorious for seeding tumor along biopsy and surgical tracts, researchers worried the percutaneous laser trajectory could do the same. That risk was mitigated with an outer access cannula that isolates the tract from direct tumor contact, and no tract seeding has been observed on follow-up imaging.

The authors are careful about what these data can and cannot show. Because radiosurgery followed four of the five procedures, the independent contribution of LITT cannot be separated from that of radiation in this retrospective design. Still, several observations hint that the laser is doing real work: the one procedure not followed by radiosurgery nonetheless showed volumetric reduction, and shrinkage appeared on scans as early as 2.8 months, a timeframe consistent with the immediate ablative effect of heat rather than the slower radiobiological response to radiation. Prior large sLITT series in metastatic disease, including one of 129 patients treated across 144 thoracic segments, found that patients treated with LITT alone achieved one-year local control of 83.4 percent, statistically similar to those who also received radiotherapy.

The researchers position sLITT as a complement to, not a replacement for, established therapies. Repeat resection remains the most definitive local treatment but carries high complication rates in previously irradiated fields, and reirradiation is constrained by cumulative dose to the spinal cord and other organs at risk. The realistic niche for sLITT is narrow but meaningful: local tumor reduction and decompression in patients whose lesions sit in previously treated fields, who need to resume systemic therapy quickly, or for whom open surgery is too risky. Intriguingly, the team also raises a forward-looking hypothesis. Chordoma harbors an immune-excluded microenvironment in which stromal barriers wall off effector immune cells, and LITT-induced hyperthermia has been shown in other tumors to release damage-associated molecular patterns and transiently open the blood-brain barrier, potentially converting immunologically cold tumors into hot ones. Whether laser ablation could help unlock chordoma’s defenses against immunotherapy remains untested, but it offers one more reason this small proof-of-concept study may mark the beginning of a much larger story.

Subject of Research: MRI-guided spinal laser interstitial thermal therapy as a minimally invasive treatment for chordoma

Article Title: Spinal laser interstitial thermal therapy for treatment of chordoma: single-center experience and literature review

Article References: Jonathan, S. V., Chen, J. W., Wu, K., Flores, A. R., Ghia, A. J., Kimchi, G., North, R. Y., Alvarez-Breckenridge, C. A., Rhines, L. D., & Tatsui, C. E. (2026). Spinal laser interstitial thermal therapy for treatment of chordoma: single-center experience and literature review. Journal of Neuro-Oncology, 179(2), Article 59. https://doi.org/10.1007/s11060-026-05764-4

Image Credits: AI Generated

DOI: 10.1007/s11060-026-05764-4

Keywords: chordoma, laser interstitial thermal therapy, LITT, spinal tumor, MRI thermometry, stereotactic radiosurgery, neurosurgery, MD Anderson, salvage therapy, thermal ablation, tumor microenvironment, Journal of Neuro-Oncology

The research was led by Jia Wu, Ph.D., associate professor of Imaging Physics and of Thoracic/Head and Neck Medical Oncology at MD Anderson and an affiliate member of the institution’s Institute for Data Science in Oncology. Co-senior authors include Ajay Sheshadri, M.D., associate professor of Pulmonary Medicine, and Mehmet Altan, M.D., associate professor of Thoracic/Head and Neck Medical Oncology. Their team set out to solve a problem that has long frustrated oncologists: immune checkpoint inhibitors have transformed outcomes in lung cancer, but the same drugs that unleash T cells against tumors can also trigger the immune system to attack healthy lung tissue, producing pneumonitis that is difficult to anticipate before symptoms appear.

“Pneumonitis remains one of the most challenging complications of immunotherapy because it can be difficult to predict before symptoms appear,” Wu said. “Our model was able to identify signals associated with future risk using information that already exists in routine CT scans.” That distinction matters clinically, because once symptomatic pneumonitis develops, patients often require high-dose corticosteroids or other immunosuppressive therapy, and in severe cases the drugs that were controlling their cancer must be stopped entirely. A method for identifying high-risk patients before treatment could allow oncologists to schedule closer monitoring, order earlier imaging follow-up, or enroll vulnerable patients in prevention trials.

The model, named the Checkpoint-Inhibitor Pneumonitis Hazard EstimatoR, or CIPHER, belongs to a class of systems known as AI foundation models. Rather than being trained narrowly on labeled examples of pneumonitis, the model was first trained on an enormous volume of chest CT data, more than 590,000 individual image slices drawn from 2,500 patients with lung cancer. During this phase it learned to recognize the underlying structural patterns of lung tissue: the fine meshwork of airways and vessels, the texture of healthy parenchyma, and the visual signatures of scarring, emphysema, and inflammation. This strategy, sometimes called self-supervised or transfer learning, lets the model build a rich internal representation of normal and abnormal lung anatomy before it is ever asked to make a clinical prediction.

Only after learning lung anatomy did the researchers ask whether the patterns CIPHER had absorbed could distinguish patients who later developed pneumonitis from those who did not. The team tested the model on pretreatment CT scans from 347 patients with non-small cell lung cancer treated at UT MD Anderson, and then validated it against an independent external dataset. In both cohorts, CIPHER achieved an area under the curve, or AUC, of approximately 0.83, a statistical measure in which 0.5 represents chance-level guessing and 1.0 represents perfect discrimination. An AUC of 0.83 indicates strong predictive performance, and critically, the model outperformed both conventional models built on clinical risk factors and radiomics approaches, which extract hand-designed quantitative features from images.

Robustness across different conditions is often the stumbling block for medical AI, and it is where CIPHER delivered some of its most persuasive results. The external validation cohort differed from the original in patient population, in the types of CT scanners used, and in imaging protocols, variables that frequently degrade the performance of image-based models when they are deployed outside their training environment. CIPHER nonetheless maintained its level of accuracy, suggesting the model captured genuine biology rather than scanner-specific artifacts or population quirks. Patients the model classified as high-risk also tended to develop pneumonitis sooner after starting immunotherapy, an indication that the model may be measuring a continuous degree of lung vulnerability rather than simply sorting patients into binary categories.

The team further tested whether CIPHER’s predictions stood up when adjusted for known clinical variables. Even after accounting for age, smoking history, tumor histology, and prior radiation therapy to the chest, factors that independently influence pneumonitis risk, the model’s predictions remained statistically significant. Prior thoracic radiation in particular is recognized as a major contributor to treatment-induced lung injury, so a model that adds predictive value on top of these established risk factors is capturing something that clinicians currently cannot see. “What makes this approach particularly interesting is that it was not designed to look for pneumonitis itself,” Wu said. “Instead, the model learned patterns within lung tissue and identified subtle abnormalities associated with future risk. That suggests routine imaging may contain much more information about treatment toxicity than we previously recognized.”

The findings arrive at a moment when immune checkpoint inhibitors have become a cornerstone of lung cancer therapy. These drugs work by releasing molecular brakes, such as the PD-1 and CTLA-4 pathways, that normally restrain T cell activity, allowing immune cells to recognize and attack tumor cells. The same mechanism, however, can produce immune-related adverse events in nearly any organ, with pneumonitis among the most feared because of its potential severity and because it can be difficult to distinguish from infection, tumor progression, or radiation injury. Current risk assessment relies on subjective radiologic review and broad clinical factors that fail to identify most of the patients who will ultimately be affected, leaving clinicians to detect pneumonitis only after cough, shortness of breath, or infiltrates on follow-up imaging appear.

The MD Anderson researchers caution that several steps remain before CIPHER can enter routine clinical use. Additional prospective studies involving larger and more diverse patient populations will be needed to determine whether the approach can be integrated into clinical workflows, where it would ideally run automatically on scans already being obtained for staging and treatment planning. The team also plans to evaluate whether the model performs similarly in other cancer types treated with immunotherapy, since checkpoint inhibitors are now used across melanoma, kidney, bladder, and many other malignancies, each with different baseline lung characteristics and treatment contexts.

Future work may also explore whether combining CIPHER’s imaging-based risk scores with other biomarkers, such as blood-based inflammatory markers or genetic signatures, can sharpen predictions further, and whether similar foundation models can forecast additional immunotherapy-related toxicities affecting other organs. If validated, such tools could identify candidates for prevention studies, guide the intensity of monitoring during treatment, and deepen scientific understanding of how immune-related side effects develop in the first place. The research was supported by the National Institutes of Health, the Cancer Prevention and Research Institute of Texas, and institutional funding from UT MD Anderson, and the study was published in the Journal for ImmunoTherapy of Cancer.

Subject of Research: An AI foundation model that predicts immune checkpoint inhibitor-induced pneumonitis risk in lung cancer patients from routine pretreatment CT scans.

Article Title: AI model uses routine imaging to identify patients at risk for serious treatment-induced lung inflammation

Article References: AI model uses routine imaging to identify patients at risk for serious treatment-induced lung inflammation. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: artificial intelligence, immunotherapy, pneumonitis, lung cancer, CT imaging, deep learning, checkpoint inhibitors, MD Anderson, risk prediction, radiomics, non-small cell lung cancer, medical imaging