Title: In vivo validation of automated CT-based volumetric assessment of lung metastases

Authors: Kuhnigk JM, Bornemann L, Dicken V, Wormanns D, Krass S, Peitgen HO

Purpose:
The assessment of metastatic growth under chemotherapy is currently performed by manual measurements of
largest nodule diameters (RECIST). Aim of the studies was to show that computer assisted 3d volumetry is
beneficial to improve speed and reliability of growth assessment in therapy response evaluation.

Method and Materials:
In vivo reproducibility studies were conducted concerning the volumetric analysis of 105 metastases
(diameters 5-50 mm) from 8 patients which were examined with a low-dose four-detector-row CT (120 kVp, 20 mAs
(effective), collimation 4x1 mm, slice thickness 1.25 mm, reconstruction increment 0.8 mm; Somatom
VolumeZoom, Siemens) twice within 10 minutes. Prototypical software developed under the R&D platform MeVisLab
was used for the evaluation, including a new automated approach for fast and reproducible volumetry of
pulmonary nodules which was explicitly developed for lung metastases which are frequently large, not
necessarily spherical, and often complexly attached to vasculature and chest wall. Apart from median volume
deviations and computation time measurements, the 95% limits of agreement between nodule volume measurements
were calculated using Bland and Altman statistics for assessing measurement agreement.

Results:
The performed studies revealed a segmentation success rate of 91.4% and an average computation time of 0.3
seconds on a 3.2 GHz PC. Median volume deviations in inter-observer (IO) and inter-scan (IS) tests were
measured as 0.1% (IO) and 4.7% (IS). Bland and Altman statistics revealed 95% limits of agreement for the
volumetric deviation of 7.1% (IO) and 26.9% (IS).

Conclusions:
The precision of in vivo volumetry of pulmonary metastases was shown to be sufficiently high to allow for
detection of clinically relevant volume change even in low dose CT scans, extending the usability of lesion
volumetry methods from screening nodules to metastases in chemotherapy monitoring. Considering that for
roughly spherical nodules a volume change of 26.9% corresponds to a change in diameter of less than 10%,
automated volumetry not only increases convenience but also performance in assessing tumor growth in therapy
response evaluation.