Background <p>Pilot studies have indicated diagnostic benefits from using dual-energy CT (DECT) for staging and follow-up of melanoma patients. The purpose of this study was to investigate the sensitivity, specificity and qualitative assessment of spectral image reconstructions for metastases in melanoma patients in a large-scale, multi-reader evaluation.</p> Methods <p>In total, 308 patients with melanoma, 95 patients with metastases and a control group of 213 patients without metastases, who underwent oncologic staging CT of the chest, abdomen and pelvis on a dual-layer dual-energy CT system (dlDECT) were retrospectively included. Conventional images (CI), color-coded iodine overlays (IO) and virtual monoenergetic images at 40&#xa0;keV (VMI<sub>40keV</sub>) were reconstructed. 6 radiologists (3 experienced with 6 to 9 years and 3 less experienced with 2 to 4 years of experience) read all cases in a CI-based session, and a session based on a combination of CI, IO and VMI<sub>40keV</sub>. Readers were asked to determine presence of metastases in specific tissues in a binary fashion and to indicate diagnostic certainty and lesion delineation on 5-point Likert scales.</p> Results <p>Sensitivity for detection of metastases in the skeletal muscle and peritoneum was significantly higher for the spectral assessment (for skeletal muscle 70% vs. 61%; for peritoneum 76% vs. 62%, both: <i>p</i> &lt; 0.05). For subcutaneous metastases, there was a significant increase in specificity (92% vs. 89%, <i>p</i> &lt; 0.05), however accompanied with a significant decrease in sensitivity (79% vs. 85%, <i>p</i> &lt; 0.05). Diagnostic certainty was rated significantly higher for spectral images than CI in all (6/6) of the assessed tissues, whereas improvements in lesion delineation were noted for the skeletal muscle, the subcutaneous tissue and the pancreas.</p> Conclusions <p>We found that in melanoma patients, the benefit of dlDECT-derived spectral reconstructions depends on the assessed tissue. While assessment of skeletal muscle and peritoneal metastases was significantly improved, low or absent iodine uptake of subcutaneous lesions led to false negatives and a consecutive decrease in sensitivity.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Spectral CT imaging for assessment of metastases in melanoma patients: multi-reader evaluation

  • Christian Nelles,
  • Philip Rauen,
  • Franziska Meyer,
  • Anja Dobrostal,
  • Pia Lena Niederau,
  • Hasan Zaytoun,
  • Mathilda Weisthoff,
  • Pascale Bernard,
  • Carola Heneweer,
  • Thomas Dratsch,
  • David Maintz,
  • Jonathan Kottlors,
  • Nicole Kreuzberg,
  • Nils Große Hokamp,
  • David Zopfs,
  • Thorsten Persigehl,
  • Simon Lennartz

摘要

Background

Pilot studies have indicated diagnostic benefits from using dual-energy CT (DECT) for staging and follow-up of melanoma patients. The purpose of this study was to investigate the sensitivity, specificity and qualitative assessment of spectral image reconstructions for metastases in melanoma patients in a large-scale, multi-reader evaluation.

Methods

In total, 308 patients with melanoma, 95 patients with metastases and a control group of 213 patients without metastases, who underwent oncologic staging CT of the chest, abdomen and pelvis on a dual-layer dual-energy CT system (dlDECT) were retrospectively included. Conventional images (CI), color-coded iodine overlays (IO) and virtual monoenergetic images at 40 keV (VMI40keV) were reconstructed. 6 radiologists (3 experienced with 6 to 9 years and 3 less experienced with 2 to 4 years of experience) read all cases in a CI-based session, and a session based on a combination of CI, IO and VMI40keV. Readers were asked to determine presence of metastases in specific tissues in a binary fashion and to indicate diagnostic certainty and lesion delineation on 5-point Likert scales.

Results

Sensitivity for detection of metastases in the skeletal muscle and peritoneum was significantly higher for the spectral assessment (for skeletal muscle 70% vs. 61%; for peritoneum 76% vs. 62%, both: p < 0.05). For subcutaneous metastases, there was a significant increase in specificity (92% vs. 89%, p < 0.05), however accompanied with a significant decrease in sensitivity (79% vs. 85%, p < 0.05). Diagnostic certainty was rated significantly higher for spectral images than CI in all (6/6) of the assessed tissues, whereas improvements in lesion delineation were noted for the skeletal muscle, the subcutaneous tissue and the pancreas.

Conclusions

We found that in melanoma patients, the benefit of dlDECT-derived spectral reconstructions depends on the assessed tissue. While assessment of skeletal muscle and peritoneal metastases was significantly improved, low or absent iodine uptake of subcutaneous lesions led to false negatives and a consecutive decrease in sensitivity.