🟡 COMPETITIVE
GE HealthCare, RadNet expand mammography AI partnership
GE HealthCare and RadNet broadened their mammography AI collaboration, adding DeepHealth second-read workflows for complex cases and additional software features.
GE is doubling down on recurring-revenue AI services rather than hardware-only sales, pressuring Hologic and Siemens to match the subscription model. Watch for concrete accuracy and workflow metrics before assuming clinical gains follow the commercial commitment.
🔵 RESEARCH
Mammography AI predicts post-neoadjuvant nodal status in breast cancer
Researchers built a deep learning model that reads digital mammograms to predict axillary lymph node status after neoadjuvant therapy in breast cancer patients.
Non-invasive nodal prediction could reshape surgical planning and reduce unnecessary axillary procedures. But single-site training risks vendor and demographic bias — multi-site validation across mammography platforms is the gating step before clinical adoption.
🔵 RESEARCH
Contrast MRI 'Avocado Sign' challenges T2 for rectal nodal staging
A 120-patient retrospective study compared a contrast-enhanced MRI feature called the Avocado Sign against T2-weighted morphological criteria for mesorectal lymph node staging.
Better nodal staging sharpens rectal cancer treatment decisions, but a brand-new sign from a single centre needs inter-reader and cross-platform validation before radiologists should change reporting practice.
⚪ INDUSTRY
Siemens to put Healthineers spinoff to shareholder vote
Siemens is giving up its majority stake in Siemens Healthineers and will seek shareholder approval to spin off the imaging and diagnostics unit.
An independent Healthineers gains capital flexibility to compete harder with GE HealthCare and Philips on AI M&A and point-of-care expansion. Service contracts and account teams should stay stable near-term, but expect a more aggressive acquirer post-separation.
🔵 RESEARCH
Ultrasound-MRI fusion radiomics refines triple-negative breast cancer prognosis
Researchers developed a deep learning radiomics nomogram fusing ultrasound and MRI data to stratify prognosis in triple-negative breast cancer patients receiving neoadjuvant chemotherapy.
Multimodal fusion strengthens the case for cross-modality AI platforms and pressures pure-play ultrasound vendors to partner or integrate with MRI players. Real deployment demands synchronized protocols and workflow redesign few breast centres have today.
Journal of Ultrasound in Medicine · Sat, 18 Ap
Read more → (paywall)
🔵 RESEARCH
Trial uses AI bladder POCUS to gate pelvic ultrasound readiness
A clinical trial is testing handheld AI-enabled POCUS to objectively measure bladder fullness and confirm patient readiness for pelvic ultrasound.
Addresses a genuine workflow friction — unprepared patients drive repeat exams — but the ROI depends on native integration with GE or Philips ultrasound stacks. A standalone tool risks adding another login rather than saving scanner time.
🩺 EDITOR’S TAKE — FOR CLINICIANS
Breast imaging is today's throughline — from GE-RadNet second-read workflows to mammography-based nodal prediction and ultrasound-MRI fusion nomograms. Track inter-reader reliability and multi-site validation before letting any of these reshape your surgical planning conversations.
📊 EDITOR’S TAKE — FOR INVESTORS
Siemens' move to cut loose Healthineers plus GE's deepening RadNet subscription play signals imaging is bifurcating into pure-play platforms and recurring-revenue AI services. Expect accelerated M&A as Healthineers gains capital flexibility to match GE's software pivot.
🏭 EDITOR’S TAKE — FOR INDUSTRY
OEMs are racing two tracks at once: corporate restructuring to unlock valuation (Siemens) and service-layer lock-in via AI partnerships (GE-RadNet). Health systems should pressure-test service continuity clauses and demand clinical outcome data before signing on to expanded AI workflow bundles.