Aims and objectives
Computer tomography (CT) is an established modality in the diagnosis and clinical management of patients with chronic liver disease [1;2] and it is described as a sensitive diagnostic tool for the assessment of morphological changes of liver [3-5].
To extend the value of image-based diagnosis,
recent studies investigated machine learning algorithms and their potential clinical application [6-9].
Building on this,
artificial neuronal networks have been employed to utilize implicit image information that might not be encompassed in dedicated human-made radiomic features [8;10].
The aims of...
Methods and materials
Retrospective analysis of 259 clinical CTs of patients with abdominal diseases.
Patients with focal liver parenchyma changes (e.g.
neoplasia or abscesses) and a history of liver surgery or liver interventions were excluded.
For all consecutive 259 patients the Child-Pugh Score was determined within the period of hospitalization.
The Child-Pugh Score is built on 5 sub-scores,
hereof 3 laboratory parameters (prothrombin time,
bilirubin,
and albumin) as well as 2 clinical parameters (ascites and encephalopathy).
The cut-off values for all parameters were defined according to Forman et...
Results
Epidemiologic,
laboratory,
and clinical characteristics are shown in table 2.In total,
11 significant radiomic imaging features were found that correlate significantly with the Child-Pugh Score in univariate analysis.
Spearman’s correlation coefficient was significant for all machine learning algorithms,
albeit strongest for the CNN.
The radiologists’ rating exhibited the strongest correlation (ρLR= 0.35,
ρRF= 0.32,
ρCNN= 0.51,
ρRP= 0.60; all p<0.001).
The accuracy of the CNN and RP was significant better as compared to the no information rate (ACCLR= 47%,
p= 0.47; ACCRF= 47%,
p= 0.38;...
Conclusion
Our most important finding is,
that CNNs can predict the Child-Pugh class,
as a surrogate for the severity of liver cirrhosis,
with a comparable precision to that of experienced radiologists (ρCNN= 0.51 and ACCCNN= 53% vs.
ρRP= 0.60 and ACCRP= 57%) based on a clinical multiphase CT.
Both conventional and radiomic analyses trail these performances in all assessed diagnostic scores.
Even though CT has been described as a valid tool to assess distinct morphological changes of liver parenchyma [12],
the value of multiphase liver-CT in...
References
Schwenzer NF,
Springer F,
Schraml C,
Stefan N,
Machann J,
Schick F (2009) Non-invasive assessment and quantification of liver steatosis by ultrasound,
computed tomography and magnetic resonance.
J Hepatol 51:433-445
Pamilo M,
Sotaniemi EA,
Suramo I,
Lahde S,
Arranto AJ (1983) Evaluation of liver steatotic and fibrous content by computerized tomography and ultrasound.
Scand J Gastroenterol 18:743-747
Kudo M,
Zheng RQ,
Kim SR et al (2008) Diagnostic accuracy of imaging for liver cirrhosis compared to histologically proven liver cirrhosis.
A multicenter collaborative study.
Intervirology 51...