The purpose of this randomized controlled trial is to compare the number of breast cancers detected among women who had their screening mammograms interpreted by artificial intelligence (AI) in combination with one or two breast radiologists to the number of breast cancers detected after standard independent double reading in BreastScreen Norway. The aims of the study is to prove that screening interpretation with AI in combination with one or two radiologists (the intervention) is non-inferior to standard interpretation procedure.
Independent double reading with consensus is the recommended practice for breast cancer screening programs in Europe. This is a time-consuming process and more than 99% of examinations are determined to have a negative final outcome in BreastScreen Norway. However, radiologist still overlook cancers, and review-studies have shown 20-25% of screen-detected and interval cancers to be visible at prior screening mammograms. In Norway and other countries, there is also a shortage of breast radiologists. Retrospective studies on AI in mammographic screening have shown promising results in the classification of cancer negative and cancer positive examinations. However, prospective studies are needed to elucidate questions concerning AI as a support or replacement for the radiologists in the interpretation process. In this randomized controlled trial women attending screening and consenting to participate in the study, will be randomized (1:1 allocation) to an intervention or a control group. The screening examination will be performed as usual in both groups. The intervention in the trial is the interpretation procedure. All examinations will be analyzed by AI. In the intervention group, examinations will be triaged according to risk of malignancy defined by an AI score. Examinations with a low AI score (1-7) will have a single reading by one radiologist and examinations with a intermediate or high AI score (8-10) will have an independent double reading by two radiologists. All examinations selected by either one or both radiologist will be discussed at consensus. Examinations with an AI raw score of \>9.8 will be discussed at consensus independent of the radiologists' interpretation. The radiologists will be blinded to whether the mammograms are interpreted by one or two radiologists. AI risk scores and AI image marking of suspicious calcifications and masses are only provided to the radiologists at consensus, not during independent reading. In the control group, two independent radiologists will interpret all screening examinations according to standard procedure, independent double reading. AI results will not be available at either independent reading or consensus, only for comparison with the intervention group in final analyses. Screen-detected cancers in the intervention and control arm will be analyzed in terms of non-inferiority.
Screening examinations will be analyzed by the AI system (Transpara, ScreenPoint Medical, Nijmegen, The Netherlands) that assigns each examination a cancer risk score 1-10. Examinations with a score 1-7 will be single read and examinations with score 8-10 will be double read. Examinations selected by one or two readers will be discussed in a consensus meeting. Examinations with an AI score of \>9.8 will be selected for the consensus meeting by default. AI risk scores and image markings are provided to the readers only in the consensus meetings. The readers will decide whether to recall the woman for further assessment in consensus (standard procedure).
Screening examinations will be assessed according to standard of care: independently double read, examinations selected by one or two readers will be discussed in a consensus meeting, and the radiologists will decide whether to recall the woman for further assessment. Screening examinations will be analyzed by the AI system retrospectively but the results will not be made available to the radiologist, only used for comparison to the intervention group in results analyses.
Cancer Registry of Norway, Norwegian Institute of Public Health
Oslo, Norway
RECRUITINGScreen-detected breast cancers
Number of breast cancers detected at screening per 1000 screening examinations.
Time frame: 36 months from study start-up
Consensus rate
Proportion of participant discussed in consensus among all participants (%).
Time frame: 24 months from study start-up
Recall rate
Proportion of women recalled for further assessment among all participants (%).
Time frame: 24 months from study start-up
Interval breast cancers
Number of interval breast cancers diagnosed per 1000 screening examinations.
Time frame: 60 months from study start-up
Prognostic and predictive tumor characteristics of screen-detected breast cancer
Distribution of tumor characteristics among the participants with screen-detected breast cancer.
Time frame: 36 months from study start-up
Prognostic and predictive tumor characteristics of interval breast cancer
Distribution of tumor characteristics among the participants with interval breast cancer.
Time frame: 60 months from study start-up
Sensitivity
Proportion of true positives among true positive and false negatives (%).
Time frame: 60 months from study start-up
Specificity
Proportion of true negatives among true negatives and false positives (%).
This platform is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional.
Study Type
INTERVENTIONAL
Allocation
RANDOMIZED
Purpose
SCREENING
Masking
DOUBLE
Enrollment
150,000
Time frame: 60 months from study start-up
Mammographic features of screen-detected breast cancer
Distribution of characteristics of tumors observed on the mammograms of participants diagnosed with screen-detected breast cancer.
Time frame: 36 months from study start-up
Time spent screen-reading
Time (minutes:seconds) spent by the individual radiologists interpreting screening mammograms
Time frame: 36 months from study start-up