1333P - What is the optimal annual interpretive volume for a radiologist reading screening mammograms?

Date 09 October 2016
Event ESMO 2016 Congress
Session Poster display
Topics Breast Cancer
Cancer Aetiology, Epidemiology, Prevention
Presenter Asli Uluturk
Citation Annals of Oncology (2016) 27 (6): 462-468. 10.1093/annonc/mdw385
Authors A. Uluturk, L. Pylkkanen, S. Deandrea, A. Bramesfeld, L. Neamtiu, Z. Saz Parkinson, M. Ambrosio, D. Lerda
  • Institute For Health And Consumer Protection, European Commission, Joint Research Centre, I-21027 - Ispra/IT



A positive association between the annual screening volume and the performance of screening radiologists has been suggested. The majority of studies show that there is little or no relation between radiologist's interpretive volume and sensitivity, but the association of interpretive volume to false-positive rate, cancer detection rate, recall rate or interval cancer rate is still unclear. As a consequence, many countries have adopted very variable minimum reading volume requirements (e.g., from 500 to 5,000 per year). The aim of the study is to evaluate whether an optimal annual interpretive volume for screening radiologists can be established.


A systematic review to assess the evidence of the association between radiologist interpretive volume and breast cancer screening performance outcomes, following standard Cochrane Collaboration methods, was carried out. The following databases were searched until November 2015: The Cochrane Database of Systematic Review, DARE, MEDLINE, EMBASE, PDQ, and McMaster Health Systems Evidence. GRADE methodology was applied.


From an initial set of 872 unique citations, 14 studies (2,935 radiologists, 32 screening services) were included. These studies retrospectively analysed information from prospective databases, except one study with a cross-sectional design. The evidence suggested an association between a higher annual screening volume and a lower false-positive rate, with an optimal performance around 1,500 to 4,000 readings per year (moderate quality evidence). Cancer detection rate appeared to be optimised at about 1,000 readings per year and in the context of high volume facilities (moderate quality evidence). Recall rate appeared to increase above 3,000 readings per year (low quality evidence). The effect on false-negative rate and interval cancer rate is unclear (very low quality evidence).


Screening radiologists should perform at least 1,500 to 4,000 screening mammograms per year to keep the performance outcomes at an acceptable level.This recommendation is provisional (due to uncertainty about the association between caseloads and the net benefits of screening), and conditional (depending on the availability of radiologist meeting the requirements).

Clinical trial identification

Legal entity responsible for the study

European Commission, Joint Research Centre


European Commission, Joint Research Centre


All authors have declared no conflicts of interest.