One Versus Up-to-5 Lesion Measurements for Response Assessment by PERCIST in Patients with Lung Cancer

Author(s):  
Soo Jin Kwon ◽  
Joo Hyun O ◽  
Ie Ryung Yoo
Cancers ◽  
2021 ◽  
Vol 13 (7) ◽  
pp. 1562
Author(s):  
Konstantinos Rounis ◽  
Marcus Skribek ◽  
Dimitrios Makrakis ◽  
Luigi De Petris ◽  
Sofia Agelaki ◽  
...  

There is a paucity of biomarkers for the prediction of intracranial (IC) outcome in immune checkpoint inhibitor (ICI)-treated non-small cell lung cancer (NSCLC) patients (pts) with brain metastases (BM). We identified 280 NSCLC pts treated with ICIs at Karolinska University Hospital, Sweden, and University Hospital of Heraklion, Greece. The inclusion criteria for response assessment were brain metastases (BM) prior to ICI administration, radiological evaluation with CT or MRI for IC response assessment, PD-1/PD-L1 inhibitors as monotherapy, and no local central nervous system (CNS) treatment modalities for ≥3 months before ICI initiation. In the IC response analysis, 33 pts were included. Non-primary (BM not present at diagnosis) BM, odds ratio (OR): 13.33 (95% CI: 1.424–124.880, p = 0.023); no previous brain radiation therapy (RT), OR: 5.49 (95% CI: 1.210–25.000, p = 0.027); and age ≥70 years, OR: 6.19 (95% CI: 1.27–30.170, p = 0.024) were associated with increased probability of IC disease progression. Two prognostic groups (immunotherapy (I-O) CNS score) were created based on the abovementioned parameters. The I-O CNS poor prognostic group B exhibited a higher probability for IC disease progression, OR: 27.50 (95% CI: 2.88–262.34, p = 0.004). Age, CNS radiotherapy before the start of ICI treatment, and primary brain metastatic disease can potentially affect the IC outcome of NSCLC pts with BM.


2019 ◽  
Vol 12 (2) ◽  
pp. 149-155 ◽  
Author(s):  
Diego Kauffmann-Guerrero ◽  
Andreas Schindler ◽  
Amanda Tufman ◽  
Zulfiya Syunyaeva ◽  
Thomas Pfluger ◽  
...  

2009 ◽  
Vol 27 (15_suppl) ◽  
pp. 7567-7567
Author(s):  
M. Endo ◽  
H. Watanabe ◽  
S. Yamamoto ◽  
N. Yamamoto ◽  
Y. Ohe ◽  
...  

7567 Background: The new RECIST ver. 1.1 was published in a special edition of the European Journal of Cancer in the first quarter of 2009 (EJC 2009;45:228). The major change involves the rules for lymph node measurement, which is to measure SHORT axis in stead of the longest diameter of lymph node. To be considered pathologically enlarged and measurable, a size of lymph node must be at least 15mm in short axis when assessed by CT scan. Lymph nodes that are at least 10mm but less than 15mm in short axis may be pathologic and can be considered as non-measurable/non-target lesions (not measured). The purpose of our study was to evaluate the response assessment using RECIST ver. 1.1. in comparison with that using ver. 1.0 in patients with advanced lung cancer. Methods: Two radiologists independently reviewed the objective tumour response of 305 patients (pts) with advanced lung cancer who had primary lesion and lymph node metastases as target lesions measured more than 10 mm in the longest diameter. According to ver. 1.1, only the short axis will be measured both at baseline and at follow-up. The response rates were calculated according to both the versions of RECIST. The tumor responses were divided into four categories (CR, PR, SD, PD), following proportion of agreement and estimation of kappa statistics between two the criteria as agreement measure. Results: The best overall responses as assessed by RECIST ver. 1.0 and ver. 1.1 are shown in the table. Out of the 108 pts with PR by ver.1.0, 8.3 % were downgraded to SD and 1.0 % was categorized as PD by ver.1.1. On the other hand, out of the 190 pts with SD by ver.1.0, 6.3 % were upgraded to PR and 8.9 % were downgraded to PD by ver.1.1. The proportion of agreement was 86.2 % (263/305, 95% CI: 75.8 - 96.6) and the kappa coefficient was 0.734 (95% CI: 0.662 - 0.806). Conclusions: No significant impact was observed for the revised lymph node measurement rules in the new RECIST ver. 1.1 on the response evaluation in pts with advanced lung cancer registered in this analysis. [Table: see text] [Table: see text]


2012 ◽  
Vol 30 (15_suppl) ◽  
pp. e13547-e13547
Author(s):  
Hubert Beaumont ◽  
Estanislao Oubel ◽  
Antoine Iannessi ◽  
Dag Wormanns

e13547 Background: Image-based biomarkers play an important role in the assessment of the response to therapy. The value of imaging biomarkers relies on their reproducibility, which depends on the reviewer and on the measuring system. This study aims at evaluating the impact of readers’ expertise and automation of measurements. Methods: A retrospective study was performed on 10 patients with at least one Non-Small Cell Lung Cancer (NSCLC) lesion, and followed over time (7 time points in average) with Computed Tomography (CT). 2 expert radiologists (ERs) and 5 imaging scientists (ISs) measured the Longest Axial Diameter (LAD) and the volume (VOL) of each lesion at each time point. ERs and ISs segmented the lesions by using a proprietary software providing semi-automatic segmentation processing with manual adjustment. ISs performed an additional session using manual segmentation tools only. From each segmentation, VOL and LAD were automatically computed. The variability of the measurements was calculated by using standard statistics. The response to treatment was assessed according to RECIST thresholds for LAD and with +/-30% thresholds for volume. The inter-reader agreement was measured trough the Kappa coefficient. Finally, the reviewing time with and without automation was analyzed. Results: The use of automated tools by ISs reduced the standard deviation of LAD difference from 10.7% to 8.4%. The inter-reader agreement improved Kappa from 0.57 to 0.68 for LAD, and from 0.52 to 0.69 for VOL. The automation reduced the reviewing time by a factor 4 with respect to the manual assessment. No significant differences in variability were found between ISs and the first ER, but significant differences were observed with respect to the second ER. Conclusions: In a RECIST context, automation improved significantly inter-reader agreement. When using volume as a biomarker, automation not only improved the inter-reader agreement, but also decreased notably the reviewing time. No evidence was found about the influence of the expertise on the volume measurement. The difference in the lesions interpretation by the experts is a relevant factor to account for.


2020 ◽  
Author(s):  
Yunxia Zhang ◽  
Jianfeng Huang ◽  
Qinzhou Zou ◽  
Jun Che ◽  
Kaihua Yang ◽  
...  

Sign in / Sign up

Export Citation Format

Share Document