Search

ObjectivesArtificial intelligence (AI) is seen as a major disrupting force in the future healthcare system. However, the assessment of the value of AI technologies is still unclear. Therefore, a multidisciplinary group of experts and patients developed a Model for ASsessing the value of AI (MAS-AI) in medical imaging. Medical imaging is chosen due to the maturity of AI in this area, ensuring a robust evidence-based model.MethodsMAS-AI was developed in three phases. First, a literature review of existing guides, evaluations, and assessments of the value of AI in the field of medical imaging. Next, we interviewed leading researchers in AI in Denmark. The third phase consisted of two workshops where decision makers, patient organizations, and researchers discussed crucial topics for evaluating AI. The multidisciplinary team revised the model between workshops according to comments.ResultsThe MAS-AI guideline consists of two steps covering nine domains and five process factors supporting the assessment. Step 1 contains a description of patients, how the AI model was developed, and initial ethical and legal considerations. In step 2, a multidisciplinary assessment of outcomes of the AI application is done for the five remaining domains: safety, clinical aspects, economics, organizational aspects, and patient aspects.ConclusionsWe have developed an health technology assessment-based framework to support the introduction of AI technologies into healthcare in medical imaging. It is essential to ensure informed and valid decisions regarding the adoption of AI with a structured process and tool. MAS-AI can help support decision making and provide greater transparency for all parties.

Tumor heterogeneity in breast cancer tumors is widely recognized but evolution of the breast cancer genome is complex and not yet understood. Characterizing the evolutionary landscape of breast tumors can provide biological understanding of tumor progression and metastasis may be important for directing effective treatments. We performed copy number profiling and mutation detection based on whole exome sequencing of primary tumors, lymph node and distant metastasis as well as on plasma circulating tumor DNA from twelve metastatic breast cancer patients. Preliminary data shows diverse patterns of tumor evolution. Although linear evolution with late dissemination of metastatic cells was common, parallel evolution with early dissemination from primary tumor to distant sites was evident in some cases. Comparing tissue biopsies with plasma samples, we could detect variants mirroring primary tumor and/or metastasis, depending on the time span between the progression steps. The genomic discordance between the different stages of tumor evolution in these patients not only emphasizes the importance of molecular profiling of metastatic tissue but also the possibilities of liquid biopsies for real-time tracking of tumor clonal evolution. Citation Format: Stephanie Kavan, Torben A Kruse, Lars Andersen, Martin Larsen, Malene G Hildebrandt, Anne-Marie Bak Jylling, Marianne Ewertz, Marianne Vogsen, Mads Thomassen. Mapping clonal evolution and tumor heterogeneity by whole exome sequencing of tissue and plasma circulating tumor DNA in metastatic breast cancer [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr P5-07-07.

Breast cancer is a spatially and temporally dynamic disease in which differently evolving genetic clones are responsible for progression and clinical outcome. We review tumor heterogeneity and clonal evolution from studies comparing primary tumors and metastasis and discuss plasma circulating tumor DNA as a powerful real-time approach for monitoring the clonal landscape of breast cancer during treatment and recurrence. We found only a few early studies exploring clonal evolution and heterogeneity through analysis of multiregional tissue biopsies of different progression steps in comparison with circulating tumor DNA (ctDNA) from blood plasma. The model of linear progression seemed to be more often reported than the model of parallel progression. The results show complex routes to metastasis, however, and plasma most often reflected metastasis more than primary tumor. The described patterns of evolution and the polyclonal nature of breast cancer have clinical consequences and should be considered during patient diagnosis and treatment selection. Current studies focusing on the relevance of clonal evolution in the clinical setting illustrate the role of liquid biopsy as a noninvasive biomarker for monitoring clonal progression and response to treatment. In the clinical setting, circulating tumor DNA may be an ideal support for tumor biopsies to characterize the genetic landscape of the metastatic disease and to improve longitudinal monitoring of disease dynamics and treatment effectiveness through detection of residual tumor after resection, relapse, or metastasis within a particular patient.

The benefit of FDG-PET/CT in follow-up of patients treated with adjuvant immunotherapy after resection of high-risk malignant melanoma (MM) is debated. This study evaluated the diagnostic accuracy and clinical impact of FDG-PET/CT for diagnosing MM recurrence during the first year after surgery.We retrospectively included 124 patients with resected high-risk MM, who received adjuvant immunotherapy and follow-up FDG-PET/CT. Clinical information and AJCC-8 stage was obtained from patients' medical records. Recurrence was verified by biopsy/progression on a subsequent scan leading to change of treatment. Non-recurrence was assumed when no metastases were observed until the subsequent follow-up scan. Incidence of recurrence, sensitivity, specificity, positive and negative predictive values (PPV and NPV) were outcome measures.Incidence rate of MM recurrence was 0.27 [95% CI 0.17-0.37] per person-year during the first-year. Recurrence was detected in 13 patients (10%) at 3-month FDG-PET/CT, in 10 patients (8.1%) at 6 months, 1 patient (0.8%) at 9 months, 3 patients (2.4%) at 12 months. The overall sensitivity, specificity, PPV, and NPV were 97% [86-99], 82% [78-86], 39% [29-50], and 99% [98-99], respectively. The PPV trended towards higher values as disease stage increased. At the 3-month scan, the majority of actions derived from positive findings were surgery or earlier expedition of the subsequent follow-up scan.The high rate of recurrence in patients with high-risk MM treated with adjuvant immunotherapy emphasizes the need for follow-up. The potential harm by a moderately low specificity reflecting a high number of false-positive results must be weighed against the benefit of early detection of recurrence.

This study aimed to determine if delayed (18)F-fluorodeoxyglucose ((18)FDG) PET/CT imaging improves quantitation of atherosclerotic plaque inflammation. Blood-pool activity can disturb the arterial (18)FDG signal. With time, blood-pool activity declines. Therefore, delayed imaging can potentially improve quantitation of vascular inflammation.40 subjects were prospectively assessed by dual-time-point PET/CT imaging at approximately 90 and 180 minutes after (18)FDG administration. For both time-points, global uptake of (18)FDG was determined in the carotid arteries and thoracic aorta by calculating the blood-pool corrected maximum standardized uptake value (cSUVMAX). A target-to-background ratio (TBR) was calculated to determine the contrast resolution at 90 and 180 minutes. Furthermore, we assessed whether the acquisition time-point affected the relation between cSUVMAX and the estimated 10-year risk for fatal cardiovascular disease (SCORE %). A significant increase in carotid cSUVMAX (23%, P.0001), carotid TBR (20%, P.0001), aortic cSUVMAX (14%, P.0001), and aortic TBR (20%, P.0001) was observed with time. At 90 minutes, cSUVMAX did not relate to SCORE %, whereas at 180 minutes significant positive relations were observed between SCORE % and carotid (τ = 0.25, P = .045) and aortic (τ = 0.33, P = .008) cSUVMAX.Delayed (18)FDG PET/CT imaging at 180 minutes improves quantitation of atherosclerotic plaque inflammation over imaging at 90 minutes. Therefore, the optimal acquisition time-point to assess atherosclerotic plaque inflammation lies beyond the advocated time-point of 90 minutes after (18)FDG administration.