Researchers evaluated whether artificial intelligence (AI) could help doctors better judge the cancer risk of indeterminate pulmonary nodules, which are small lung spots seen on a CT scan whose cause is not yet clear. These nodules are becoming more common as lung cancer screening and CT imaging are used more often. The team built a decision model to compare two approaches: clinician evaluation alone versus clinician evaluation supported by AI-based radiomics, a method that analyzes patterns in imaging data. They asked whether the AI approach would improve health outcomes and whether it would be worth the extra cost from a payer’s perspective over a patient’s lifetime. In their base case—a 60-year-old patient with a 1.1 cm nodule and a fairly high chance of cancer (65%)—AI support led to a small gain of 0.03 life-years and was cost-effective, with an incremental cost-effectiveness ratio of $4,485 per life-year gained. However, when the chance that the nodule was cancer was very low, below 5%, the AI approach no longer met a typical cost-effectiveness threshold of $100,000 per life-year gained. Overall, the study suggests that AI-assisted nodule assessment is cost-effective in settings where the likelihood of cancer is greater than 5%.

Fig 1.��Decision Model Structure.

The decision tree structure models the risk-stratification of an indeterminate pulmonary nodule utilizing artificial intelligence-assistance compared to the clinician alone. Repeated portions of the model have been collapsed into subtrees (A-G) for readability, each of which represents a diagnostic or management pathway that appears in various parts of the model (‘A’ = surveillance; ‘B’ = PET-CT evaluation; C = minimally invasive surgical (MIS) lobectomy; ‘D’ = low-risk surveillance; ‘E’ = intermediate-risk; ‘F’ = MIS wedge resection; ‘G’ = initial risk classification).

Advances in imaging now allow us to study bone and dental diseases in both living patients and ancient remains. In archaeological studies, dental conditions like cavities (caries) and gum disease (periodontal disease) are common and can give clues about overall health, including risks for conditions such as heart disease. Studying mummies provides a unique window into the health of past populations and the relationship between oral and systemic health over time.

Using CT scans of Egyptian mummies, we examined dental problems, including cavities, periapical lesions (infections at the tooth root), and the distance from the tooth to the jawbone. We also looked at whether oral health problems were linked to hardened arteries. We found that severe cavities and gum disease were very common in this group. Oral health issues were associated with the number of calcified blood vessels, although age and sex also influenced these patterns.

These findings show that dental disease, including cavities and gum disease, existed in ancient populations and that poor oral health may have been connected to cardiovascular disease. Understanding these historical patterns highlights the importance of oral health across the lifespan and offers context for modern dental care, emphasizing how cultural and lifestyle factors shape oral health challenges.

Figure 1��Computed tomographic scans of Hatiay, a mummy included in the original Horus work,��with evidence of both carotid calcifications (blue arrow) and poor dental health (red arrows).

Understanding how the pancreas normally changes in size and shape from infancy through old age is important for recognizing abnormal changes linked to type 2 diabetes and other pancreatic diseases. In this study, we measured pancreas morphology (size and shape) across the lifespan, from birth to age 90. Our goals were to identify reliable imaging methods for artificial intelligence (AI)-based pancreas measurement, establish normal aging patterns, and determine how type 2 diabetes may alter these patterns.

We analyzed abdominal computed tomography (CT) and magnetic resonance imaging (MRI) scans from 2,533 patients who did not have cancer, pancreatic disease, sepsis, or trauma. The scans were standardized to the same resolution, and the pancreas was automatically segmented using AI-based methods. We then extracted 13 morphological features of the pancreas.

First, we compared pancreas volume trends across contrast CT, non-contrast CT, and MRI in 1,858 control patients to determine which imaging method produced the most consistent lifespan patterns. CT was selected for the main analyses because MRI measurements differed when processed with our AI method in this clinical dataset. Next, we established normative aging patterns in pancreas morphology by age group and sex. Finally, we used statistical modeling (GAMLSS regression) to compare 675 patients with type 2 diabetes to 675 age- and sex-matched non-diabetic controls.

After adjusting for other factors, 10 of the 13 morphological features showed significantly different aging trends in people with type 2 diabetes compared to controls. Overall, the pancreas was smaller in individuals with type 2 diabetes, confirming previous findings. This study provides a large clinical reference of normal pancreas morphology across the lifespan and shows how type 2 diabetes is associated with measurable changes in pancreas size and shape.

FIGURE 1

The pancreas undergoes structural changes with age, including atrophy and fat infiltration. While population-level pancreas volume and fat content have been examined across the aging process (Saisho et��al.��), there remains a knowledge gap in understanding age-related changes in the pancreas across a broader set of morphological measurements. Moreover, type 2 diabetes may cause changes in pancreas morphology that differ from normal aging. The two scans on the left illustrate age-related appearance differences in non-diabetic patients but are not from the same patient. The rightmost scan shows the pancreas from an elderly patient with type 2 diabetes. Understanding pancreas variation in normal aging is critical for understanding differences in type 2 diabetes. Any potential differences in the aging trends of the pancreas in type 2 diabetes may not necessarily be linear or smooth.

Medical imaging, such as MRI and CT scans, has a notable environmental footprint due to energy use, equipment production, and disposable supplies. This study evaluated the environmental impact of MRI and CT services at a large academic medical center in the Southeastern United States over one year using life cycle assessment methods. Researchers collected data from direct observation, records, staff interviews, and energy metering, and assessed impacts with established environmental databases and software.

Results showed that MRI and CT services produced an estimated 221 and 108 tons of carbon dioxide equivalent annually—comparable to the emissions of 52 and 25 cars driven for a year, respectively. Energy use contributed most to emissions (58% for MRI, 33% for CT), followed by disposable supplies, equipment production, and linens. Switching to solar power could cut MRI emissions by 70% and CT emissions by 40%, though the relative contribution of supplies and equipment would then become more significant.

These findings highlight the importance of energy consumption in imaging services and suggest that renewable energy adoption, efficient scanner use, reusable supplies, and circular business practices—such as extending equipment life—can meaningfully reduce the environmental impact of medical imaging.

Fig. 1��Flow diagram of components included in the study. ∗This study could not account for the production of all additional capital equipment. See e-only����here and in the previous publication [] for more information

CT scans can look noticeably different depending on the��reconstruction kernel��used to process the images. These kernels change how sharp or noisy an image appears, which can lead to big differences in important measurements—such as how much emphysema is present in the lungs. While it’s fairly easy to make images consistent when they come from the same type of scanner, this becomes much harder in studies that collect scans from many hospitals and manufacturers. Because each manufacturer uses different kernels, the measurements can become inconsistent, making it difficult to compare results. To fix this, we need a way to standardize all CT images so they look as if they were created using the same reference kernel.

In this study, we tested whether we could train a computer model to do this standardization using both��paired��data (scans from the same person processed with two different kernels from one manufacturer) and��unpaired��data (scans from different people and different manufacturers). Our goal was to use both types of data to create a shared representation of the images that allows for consistent comparisons across all scanners.

We created a deep learning model called a multipath cycleGAN, which can learn how to “translate” CT images from one kernel style to another. It uses a shared latent space (a common internal representation), along with several encoder–decoder pathways and discriminators that help the model learn from both paired and unpaired examples. We trained the model using CT scans from seven common reconstruction kernels from the National Lung Screening Trial, giving us 42 possible kernel combinations to harmonize.

We then tested the model using hundreds of additional scans. For paired kernels, we looked at whether the model reduced differences in��percent emphysema, and it did—performing better than comparison methods in several cases. For unpaired kernels, we converted all scans to look like they were processed with a reference soft kernel or a reference hard kernel and again measured emphysema levels. Our model reduced differences in many kernel types and performed similarly or better than existing approaches. We also checked whether harmonization preserved important anatomical structures such as lung vessels, muscle, and fat, and found that our method generally maintained these details.

Overall, our results show that combining paired and unpaired data in a shared latent space multipath cycleGAN can reduce errors in emphysema measurement and keep anatomical structures consistent. This approach offers a promising way to make CT scans from different scanners and reconstruction kernels more comparable, which is important for large research studies and long-term patient monitoring.

FIGURE 1

Reconstruction kernels influence the noise and resolution of the underlying anatomical structure in a computed tomography image. (a) Paired reconstruction kernels obtained from a given vendor exhibit a one-to-one pixel correspondence between the scans, which enables kernel harmonization. However, (b) across vendors, unpaired kernels show differences in anatomy, scan protocol, field of view, and reconstruction window. This creates additional difficulties that make harmonization a more challenging task.

As climate change becomes more serious, it’s important to consider how healthcare, including cancer care, affects the environment. Imaging tests and image-guided procedures—such as CT scans, MRIs, and targeted treatments—play a major role in diagnosing and managing cancer. These technologies have improved patient outcomes, but they also produce carbon emissions and medical waste. This review looks at how cancer imaging and related treatments impact the environment, discusses what sustainable cancer care could look like, and suggests practical ways to reduce the environmental footprint of cancer care while continuing to provide high-quality treatment.

Figure 1:��Sources of greenhouse gas (GHG) emissions in cancer imaging categorized by scope. Scope 1 includes direct emissions, those from sources that an organization owns or controls directly. Scope 2 includes indirect emissions, those that come from where an organization’s energy is produced. Scope 3 includes all sources not covered in scope 1 or 2, including those created by an organization’s value chain.

To protect privacy, researchers often “deface” magnetic resonance imaging (MRI) scans of the head before sharing them publicly. This process removes or alters the parts of the scan that show a person’s facial features. However, there is ongoing debate about how well this method actually protects privacy and how much it interferes with research that depends on detailed head or brain anatomy. With advances in deep generative models, it has become unclear whether defacing truly prevents faces from being reconstructed or identified from altered MRI data.

This study developed a “refacing” pipeline that can reconstruct faces from defaced MRI scans using cascaded diffusion probabilistic models (DPMs), a type of advanced deep learning model capable of generating realistic images. The models were trained on data from 180 subjects and tested on 484 previously unseen subjects, 469 of whom came from a different dataset. The study also examined whether the altered voxel data—the tiny 3D units that make up MRI images—still contain useful anatomical information. Specifically, it tested whether computed tomography (CT)-derived skeletal muscle radiodensity could be predicted from facial voxels in both defaced and original MRIs.

The results show that the DPMs could reconstruct faces that closely resembled the originals, with surface distances significantly smaller (p < 0.05) than those between the original faces and population-average faces. This indicates that defacing may not be sufficient to guarantee privacy, as realistic facial features can be recovered even after defacing. The refacing performance also generalized well to new datasets, meaning the technique worked consistently beyond the data it was trained on.

When it came to predicting muscle radiodensity, defacing was found to reduce accuracy. Using defaced images led to significantly lower Spearman’s rank correlation coefficients (p ≤ 10⁻⁴) compared to using original scans. For the shin muscle, predictions were statistically significant (p < 0.05) when using original images but not significant (p > 0.05) with any defacing method. These results suggest that defacing may not only fail to protect privacy but also remove valuable anatomical information that could aid medical research.

The study proposes two potential solutions to balance privacy and scientific value: first, sharing skull-stripped images (with the facial and cranial regions removed) along with measurements of those regions taken beforehand, though this limits research possibilities; and second, sharing unaltered images but enforcing privacy through strict data-use policies rather than through image alteration.

Fig. 1.

��There are pitfalls of defacing, a technique used to alter facial voxels in whole-head MRIs to protect privacy. First, with deep generative models such as diffusion probabilistic models, it is possible to synthesize MRIs with realistic faces, which closely resemble the original faces, from defaced MRIs. This capability poses a re-identification risk, thus questioning the efficacy of defacing in protecting privacy. Second, facial and other non-brain voxels in whole-head MRIs contain valuable anatomical information. For instance, this information could be used to study correlations between head and body measurements using paired head MRI and body CT data. The alteration of these voxels results in information loss, thereby compromising such research potentials. The experiments in this paper are designed to showcase these two pitfalls.

There is growing interest in how long cancer has existed and why malignant tumors, especially in soft tissues, seem rare in ancient human remains. To explore this, researchers carefully examined 45 whole-body CT scans of ancient Egyptian mummies to look for bone and soft tissue tumors. They found evidence of malignant bone disease (likely cancer spread to the bones) in 1 out of 45 cases (2%). In addition, 5 out of 45 mummies (11%) showed soft tissue masses that were likely cancerous. These soft tissue tumors had clear edges, different internal patterns, and were denser than the nearby preserved soft tissues. In two cases where soft tissue tumors were inside the abdomen, the original organs were not preserved. In summary, malignant tumors, including those in soft tissues, can be detected using CT scans of ancient Egyptian mummies. This discovery about how these tumors appear and how often they occur provides new information and a fresh way to study cancer in ancient populations.

Fig 1

Case 16, probable skeletal metastases and large intra-abdominal soft tissue mass. (A) Axial multiplanar (MPR) reconstruction of the skull illustrating multiple predominantly small osteolytic lesions of the cranial vault involving the outer and inner table as well as the diploe. (B) Sagittal MPR of the cervical and thoracic spine demonstrating multiple osteolytic lesions in the vertebral bodies, spinous processes and the sternum. The second thoracic vertebra reveals a healed burst fracture with collapse, the first and third thoracic vertebrae show infraction of the adjacent endplates. (C) Sagittal MPR of the lumbar spine and the sacrum showing multiple osteolytic lesions. Note the cachectic body and the textiles that overly the lumbar spine and protrude into the relatively empty pelvic cavity (asterisk). (D) Axial MPR of the upper abdomen illustrates a large soft tissue mass that expands from the midline towards the left dorsolateral part of the abdomen (arrows). In the midline, the mass is relatively homogeneous, in the lateral part, it shows different components/layers with stratified, loosened structure. The mass appears to remodel the pancreas body and tail. (E) Axial MPR of the middle abdomen demonstrating the lower part of the mass with different components/layers (arrows). (F) Coronal MPR of the upper abdomen illustrating the soft tissues mass with irregular contour of the upper margin (arrows). Note preservation of shrunken lungs bilaterally. Unfortunately, the CT scan was sectioned at the level of the soft tissue mass. (G) Coronal MPR of the middle and lower abdomen showing the mass with a horizontal postmortem split due to desiccation (arrow). There is increased adjacent soft tissue around the lower margin (asterisk).

As more older adults become candidates for liver transplants, it’s increasingly important to distinguish between dementia caused by aging and��hepatic encephalopathy (HE)—a condition related to liver dysfunction. These two conditions can appear very similar in how they affect thinking and memory. While��neuropsychological testing (NPT)��is commonly used to evaluate dementia, imaging techniques like��FDG PET/CT scans, which measure brain activity, are being explored as faster alternatives. This preliminary study looks at how useful FDG PET/CT is for detecting irreversible dementia in older people being considered for a liver transplant.��

Eighteen patients showing signs of possible dementia during their transplant evaluations received an FDG PET/CT brain scan. The study compared the PET/CT scan results to NPT results and also looked at how long each test took to produce a diagnosis. Brain scan results were also compared to healthy individuals of the same age to see if there were noticeable differences in brain activity.��

In 40% of the patients, the PET/CT scan identified signs of irreversible dementia. The scan results moderately agreed with the NPT results, and while the scan was highly sensitive (able to catch all cases of dementia), it wasn’t as specific (sometimes showing false positives). Importantly, the PET/CT scan delivered results much faster—on average in about 12 days, compared to over 70 days for NPT. The scans also revealed significant reductions in brain activity in areas linked to thinking and movement, compared to healthy individuals.��

These findings suggest that��FDG PET/CT scans could be a useful early tool��in identifying irreversible causes of dementia in older liver transplant patients. However, a positive scan should not be the sole reason to rule someone out of receiving a transplant; it should lead to further testing. Using PET/CT scans as an initial screening step could speed up the evaluation process and help ensure that patients who need further care get it in time.��

��

Fig. 1��

Results from matched case-control comparison of FDG PET scans of the brain. Cases (patients referred for FDG PET for cognitive symptoms during transplant evaluation) had significantly decreased fluorodeoxyglucose (FDG) uptake in multiple large clusters (outlined in white) primarily in the frontal, temporal, and parietal lobes compared to healthy age- and sex-matched controls from the Alzheimer’s Disease Neuroimaging Initiative database (ADNI). These T-statistic maps show the regional statistical differences in glucose metabolism in the brain between 16 case-control pairs. Blue reflects voxels where cases had decreased FDG uptake vs. controls, while red reflects the opposite. Clusters that reach significance (FDR < 0.05 and voxel-extent of 10) are outlined in white. No significant clusters of increased SUVR were observed.

��

Contrast-enhanced CT scans��use special dyes (called radiocontrast agents) that highlight blood vessels by making them appear brighter than the surrounding tissue. To get the best images, it’s important to time the scan correctly—right when the contrast is at its strongest in the area being examined.��

This study introduces a new method for predicting the��optimal timing of contrast��during a CT scan using a type of machine learning called a��circular regression model. Instead of treating contrast timing as one of a few fixed stages (as many previous methods do), this technique treats timing as a��continuous��value. That allows for more precise predictions and better adjustment to differences between patients—especially how contrast flows through each person’s blood vessels.��

The model uses��2D convolutional neural networks��(a kind of AI that processes image data) to learn patterns from earlier time points in a scan and predict the best contrast timing. It was trained on 877 CT scans and tested on 112 new scans, achieving��93.8% accuracy, which is on par with the best current methods. The results show that this new approach, which focuses on prediction rather than classification, performs better than existing 2D and 3D models that try to label scans into fixed categories.��

The study also investigates how the��location of each CT slice in the body��relates to contrast timing, suggesting that this information could help make predictions even more accurate—a new idea that hasn’t been explored before.��

��

Fig. 1.����

Four contrast phases used as anchor points in our re-gression model. Organs (e.g. kidneys) are enhanced differently in each phase. The difference between some phases are subtle, e.g., between Nand D, making automatic identification challenging.��

��