AJ Watson

[email protected]

The cloud for healthcare providers: One size does not fit all

In a new column, AJ Watson of Intelerad Medical Systems discusses the merits of a hybrid approach to cloud computing for medical image management.

July 23, 2023

2020 01 28 20 45 5229 Cloud Computing Laptop 400

What's next for AI regulations in medical imaging?

The U.S. Food and Drug Administration continues to consider its approach to regulating artificial intelligence (AI) software. AJ Watson of Intelerad Medical Systems makes the case for flexible and adaptable regulation of the technology.

September 26, 2022

2019 12 09 17 23 6730 Artificial Intelligence Ai Suit 400

Page 1 of 1

Cropped secondary capture examples. These are illustrative and not intended to imply superiority or inferiority of any device. (A) Posteroanterior radiograph in a 46-year-old female patient. The device correctly identified a right lower lobe nodule projected below the right hemidiaphragm and hilar lymphadenopathy. (B) Posteroanterior radiograph in an 86-year-old female patient with a classic Golden S sign highly suggestive of cancer. Three devices did not identify any findings. (C) The output from one device for the same radiograph as in B. The device placed a contour around the area of abnormality but mislabeled it as segmental collapse, and there are no other elements in the output to raise suspicion of cancer. (D) Posteroanterior radiograph in a 60-year-old male patient -- a case of confirmed lung cancer that was not deemed visible in retrospect. The device identified multiple false-positive abnormalities. (E) Posteroanterior radiograph in a 77-year-old female patient with two right lower lobe nodules. The device mislabeled the abnormality as infection -- a diagnostic term that could incorrectly influence clinical management. (F) Posteroanterior radiograph in a 77-year-old female patient with a right hilar tumor. Most of the lungs have been labeled by the device, with excessive overlap of the abnormality that pragmatically represents an incorrect result. All annotations shown were produced by the devices. LL = lung lesion, LO = lung opacity, PO = pleural other, TBC = tuberculosis.

AI devices vary widely in lung cancer detection

Three devices helped detect more cancerous tumors, whereas the other four devices helped detect fewer tumors.

Overview of the study design, including data collection, radiomics model construction, and biologic interpretation. (A) The primary cohort for model development and evaluation in the training and internal test sets includes 168 patients with colorectal cancer (CRC) from center 1 (Fudan University Shanghai Cancer Center), whereas external testing was performed using data from 85 patients from centers 2 (Huashan Hospital, Fudan University) and 3 (Zhongshan Hospital, Fudan University). For model interpretation, RNA sequencing (RNA-seq) data from 18 patients with CRC in center 1 and single-cell data from 35 patients from the Gene Expression Omnibus (GEO) database are incorporated. Mismatch repair (MMR) protein expression, obtained from pathologic reports in electronic medical records (EMRs), was used to identify patients with consensus molecular subtype 1 (CMS1) CRC. Tissue microarrays (TMAs) from surgically obtained CRC tissues were analyzed for immunohistochemical staining of caudal-type homeobox 2 (CDX2), 5-hydroxytryptamine receptor 2B (HTR2B), FERM domain containing 6 (FRMD6), and zinc finger E-box binding homeobox 1 (ZEB1), and the results were subsequently entered into an online classification tool that assigned patients to CMS2, 3, or 4 subtypes. (B) Radiomics workflow for CMS4 prediction. Lesions were manually delineated on T2-weighted imaging (T2WI) and contrast-enhanced (CE) T1-weighted imaging (T1WI), followed by radiomics feature extraction. Feature selection and model development were performed in the training set using machine-learning algorithms with fivefold cross-validation for model selection, and the final model was evaluated in the internal and external test sets using receiver operating characteristic (ROC) curve analysis. (C) Transcriptomic interpretation workflow. Using the radiomics-predicted CMS4 classification, bulk RNA sequencing (RNA-seq) data were used to perform differential expression analysis between predicted CMS4 and non-CMS4 groups, followed by pathway enrichment analyses to assess biologic relevance. Public single-cell RNA sequencing data from the Gene Expression Omnibus database were then used for cell-type–level interrogation to support interpretation of the associated pathways.

MRI-based ML model shows promise in colorectal cancer subtyping

The Invisible Force Podcast Album Cover Auntminnie Jan 2026 Thumbnail sra O0 D Ij Z6

Podcast: Expert witness hats come out to begin closing investigation

Mri Scanner Kid

Almost half of families face financial hardship from kids' scans

Radiology is Under Pressure

Radiology is Under Pressure

AI devices vary widely in lung cancer detection

Three devices helped detect more cancerous tumors, whereas the other four devices helped detect fewer tumors.

May 19, 2026

Cropped secondary capture examples. These are illustrative and not intended to imply superiority or inferiority of any device. (A) Posteroanterior radiograph in a 46-year-old female patient. The device correctly identified a right lower lobe nodule projected below the right hemidiaphragm and hilar lymphadenopathy. (B) Posteroanterior radiograph in an 86-year-old female patient with a classic Golden S sign highly suggestive of cancer. Three devices did not identify any findings. (C) The output from one device for the same radiograph as in B. The device placed a contour around the area of abnormality but mislabeled it as segmental collapse, and there are no other elements in the output to raise suspicion of cancer. (D) Posteroanterior radiograph in a 60-year-old male patient -- a case of confirmed lung cancer that was not deemed visible in retrospect. The device identified multiple false-positive abnormalities. (E) Posteroanterior radiograph in a 77-year-old female patient with two right lower lobe nodules. The device mislabeled the abnormality as infection -- a diagnostic term that could incorrectly influence clinical management. (F) Posteroanterior radiograph in a 77-year-old female patient with a right hilar tumor. Most of the lungs have been labeled by the device, with excessive overlap of the abnormality that pragmatically represents an incorrect result. All annotations shown were produced by the devices. LL = lung lesion, LO = lung opacity, PO = pleural other, TBC = tuberculosis.

MRI-based ML model shows promise in colorectal cancer subtyping

By Amerigo Allegretto

A machine-learning model based on MRI radiomics achieved high performance in predicting colorectal cancer subtypes.

May 19, 2026

Overview of the study design, including data collection, radiomics model construction, and biologic interpretation. (A) The primary cohort for model development and evaluation in the training and internal test sets includes 168 patients with colorectal cancer (CRC) from center 1 (Fudan University Shanghai Cancer Center), whereas external testing was performed using data from 85 patients from centers 2 (Huashan Hospital, Fudan University) and 3 (Zhongshan Hospital, Fudan University). For model interpretation, RNA sequencing (RNA-seq) data from 18 patients with CRC in center 1 and single-cell data from 35 patients from the Gene Expression Omnibus (GEO) database are incorporated. Mismatch repair (MMR) protein expression, obtained from pathologic reports in electronic medical records (EMRs), was used to identify patients with consensus molecular subtype 1 (CMS1) CRC. Tissue microarrays (TMAs) from surgically obtained CRC tissues were analyzed for immunohistochemical staining of caudal-type homeobox 2 (CDX2), 5-hydroxytryptamine receptor 2B (HTR2B), FERM domain containing 6 (FRMD6), and zinc finger E-box binding homeobox 1 (ZEB1), and the results were subsequently entered into an online classification tool that assigned patients to CMS2, 3, or 4 subtypes. (B) Radiomics workflow for CMS4 prediction. Lesions were manually delineated on T2-weighted imaging (T2WI) and contrast-enhanced (CE) T1-weighted imaging (T1WI), followed by radiomics feature extraction. Feature selection and model development were performed in the training set using machine-learning algorithms with fivefold cross-validation for model selection, and the final model was evaluated in the internal and external test sets using receiver operating characteristic (ROC) curve analysis. (C) Transcriptomic interpretation workflow. Using the radiomics-predicted CMS4 classification, bulk RNA sequencing (RNA-seq) data were used to perform differential expression analysis between predicted CMS4 and non-CMS4 groups, followed by pathway enrichment analyses to assess biologic relevance. Public single-cell RNA sequencing data from the Gene Expression Omnibus database were then used for cell-type–level interrogation to support interpretation of the associated pathways.

Podcast: Expert witness hats come out to begin closing investigation

The episode the hosts have eagerly anticipated.

May 19, 2026

The Invisible Force Podcast Album Cover Auntminnie Jan 2026 Thumbnail sra O0 D Ij Z6

Almost half of families face financial hardship from kids' scans

By Kate Madden Yee

These families are more than three times as likely to delay or skip recommended scans because of cost.

May 19, 2026

Mri Scanner Kid

Combined fluoride and lead exposure reduces bone health in youth

Co-exposure to fluoride and lead exacerbates changes in BMD more than single exposure, suggesting synergistic effects on bone impairment.

May 18, 2026

A graphical abstract of the study.

Medical societies applaud proposed CEUS training recommendations

By Amerigo Allegretto

Ultrasound societies are voicing support for new proposed training standards by CAAHEP for CEUS education.

May 18, 2026

Contrast Injector 400

Cochrane updates 2013 prostate cancer screening review

By Kate Madden Yee

Standard PSA screening has a well-known weakness: it catches many slow-growing cancers that would never have caused harm, Cochrane said.

May 18, 2026

ISMRM: fMRI flags brain network failures in fibromyalgia, long COVID

By Kate Madden Yee

While both conditions demonstrate reduced brain network flexibility, the underlying mechanisms differ.

May 18, 2026

Muscle Trigger Points Sized

ISMRM: PET/MRI highly effective in patients with atherosclerosis

PET/MRI distinguished symptomatic from asymptomatic carotid plaques among 145 patients.

May 15, 2026

Multimodality imaging (MRI, PET, and fused PET/MRI) of carotid artery plaques in asymptomatic and symptomatic subjects, illustrating distinct features of surface ulceration, plaque enhancement, metabolic uptake, and SUVmax.

ISMRM: Researchers eye fibrosis patterns in the body's organs

Quantitative measures obtained from MRIs enable fibrotic scoring system, according to group.

May 15, 2026

Fibrotic Buckets Disease Inclusion Uk Biobank Population Ismrm2026

ISMRM: PET/MRI-guided treatment reduces low back pain

This study is the first to acquire follow-up FDG-PET/MRI scans after treatment to assess the correlation between clinical outcome and FDG uptake.

May 15, 2026

A 70-year-old female patient with chronic low back pain. History of radicular syndrome which led to targeted nerve block with positive effect. However, lower back pain persisted. FDG-PET/MRI showed increased FDG uptake in a disc herniation at L5-S1 (SUVmax 1.6) convincing the treating physician to proceed to discectomy. After discectomy, the patient is now mostly pain free.

ISMRM: MRI-based AI predicts organ aging before disease signs emerge

By Kate Madden Yee

A DL model combining MRI scans with blood markers, imaging biomarkers, and lifestyle data can predict organ-specific biological age.

May 15, 2026

Ismrm Thumbnail

Page 1 of 225