New workflow method for ultrasound examinations

Sammanfattning: Background: Growing demand for ultrasound examinations and higher quality requirements motivate searching for routines combining the diagnostic accuracy of radiologist-performed examinations with the economical advantages of sonographerperformed examinations. One possible approach is to use strictly standardized examination protocols and documentation made by cine-loops that will give the radiologist access to all relevant information after the examination. Ultrasound examinations are usually regarded as observer dependent, but using documentation with cine-loops acquired in a standardized way attempts to reduce this problem.Aims: The aim of study I was to compare a recently introduced routine, combining acquisition by a radiographer, documentation as standardized cine-loops, and review by a radiologist (“standardized method”), with the formerly used routine where the diagnosis is made bed-side by the radiologist (“traditional method”). The aim of study II was to evaluate the intra-observer and inter-observer agreement of the standardized method in ultrasound liver examinations.Material and Methods: In study I there was 64 policlinic patients examining the kidneys ( n = 27) or the gallbladder ( n =37) by both the standardized and the traditional method. The radiologists’ findings of hydronephrosis, tumors, cysts, echogenicity changes, and cortical thickness (in the kidneys), and wall thickness, concrements, and polyps (in the gallbladder) were compared between the methods with respect to agreement as well as systematic differences. In study II 98 out-patients were examined by a radiographer using the standardized method. Three radiologists with 10 – 20 years of experience of ultrasound reviewed the cine-loops retrospectively and independently filled out a predetermined protocol. After 4 weeks, the review was repeated, blinded to the initial reading.Results: Study I showed for the gallbladder examination a median agreement of 97% (86 – 100%; kappa =0.64 – 1.00) and for the kidney examination an agreement of 90% (78 – 100%; kappa = 0.69 – 1.00). There were no significant systematic differences between the two methods. In study II, the intra-observer agreement was highest for concrements in the gallbladder (kappa = 0.91 to 0.96) and lowest when assessing the need for further examination (kappa = 0.38 to 0.64). For increased liver echogenicity, kappa varied between 0.51 and 0.85, and for skip areas between 0.73 and 0.90. The interobserver agreement was also highest for concrements in the gallbladder (kappa = 0.84 to 1.00) and lowest for need for further examination (kappa = -0.12 to 0.46). For most other findings, substantial intra-observer agreement (kappa ≥ 0.61) was found.Conclusion: The satisfactory agreement in study I indicates that the new workflow with ultrasound examinations performed by a radiographer and analyzed off-line by a radiologist is promising. Study II shows a fairly good inter-observer agreement for ultrasound examinations acquired with a standardized technique by an experienced radiographer and reviewed by experienced radiologists. In general, intra-observer agreement was higher than inter-observer agreement.