This study was carried out on 21 radiographer students (17 women, 4 men, mean age: 21.3 years). These 3rd year students have already received all the theoretical training related to the post-processing protocols and have all accumulated at least 8 weeks of full-time internships in a CT department (Fig 2).
Each post-processing protocol was performed 15 times by each student over a 4-week period in the training centre (Fig 2 and Fig 3). The post-processing protocols were carried out on Horos software (v2.1.1) (horosproject.org) on identical computers (iMac 21.5 inches, Apple, cupertino, USA) (processor: 1.4 Ghz intel Core i5 dual core, RAM: 16GB DDR3 1600 MHz, graphics card intel HD Graphics 5000).
Protocol #1 consists of performing MPR and VR reconstructions from a CT of a left shoulder, in a traumatic context (Fig 4).
Protocol #2 consists of performing MPR and MIP reconstructions from a CT of a middle ear, in the case of otosclerosis (Fig 5). Each post-processing was timed using the Timer application (V 7.0.0.0, Apimac, Padova, Italy) installed on each station.
Each repetition (n=15) for each protocol (n=2) was timed for each student (n=21).
The data obtained were evaluated using a logarithmic approach (Wright model) (1). To do this we correlated our results with the following function:
TN = T1 · Nb
With
TN = the duration of the Nth post-processing operation
T1 = the duration of the 1st post-processing operation
N = the operation number (post-processing)
b = (log of the learning rate)/(log 2) = coefficient of the learning curve
For each protocol (n=2) and for each repetition (n=15)) the accuracy of the work accomplished was evaluated (Fig 3) by an experienced trainer by proposing a binary score scale. If all criteria were validated, the score was 1; if one or more criteria were not validated, the score was 0 (work performed was considered non-exploitable). For each repetition (and for each protocol) the average score (between 0 and 1) was calculated. The data obtained were correlated using ProFit software (V7.0, Quantomsoft, Uetikon am See, Switzerland).