Modeling of mesothelioma growth demonstrates weaknesses of current response criteria☆
Introduction
Malignant mesothelioma is a deadly malignancy affecting an estimated 2500 Americans each year. Due to advanced disease at presentation, systemic chemotherapy is the only treatment option for the majority of these patients [1]. Active agents include pemetrexed [2], gemcitabine [3], and vinorelbine [4], and many promising new agents are currently under investigation [5]. The development of treatments that improve symptoms and prolong survival in mesothelioma patients emphasizes the need for accurate tumor measurement techniques for the assessment of response to therapy.
Radiologic assessment of change in tumor size has become a standard way to evaluate the efficacy of chemotherapeutic agents in solid tumors such as mesothelioma [6]. An early standardization of tumor response assessment was published in 1981 by the World Health Organization (WHO) [7]. These guidelines comprised two elements: (1) a tumor measurement technique and (2) a set of corresponding response criteria. The WHO measurement technique requires acquisition of the tumor's longest diameter and the corresponding longest perpendicular diameter, the product of which constitutes a bidimensional tumor measurement. The WHO response criteria define partial response (PR) as a ≥50% decrease over time in the bidimensional measurements summed over all measurable lesions and progressive disease (PD) as a ≥25% increase in summed measurements. These criteria for PR and PD will be denoted [−50%, +25%].
The WHO guidelines were used widely until a technique based solely on unidimensional measurement of a tumor's largest dimension was proposed [8]. This approach exploits the consistent mathematical relationships among the diameter, cross-sectional area, and volume of a spherical tumor [9]. The Response Evaluation Criteria in Solid Tumors (RECIST) [10] guidelines dictated a unidimensional tumor measurement technique with corresponding response criteria that defined PR as a ≥30% decrease in the sum of the unidimensional measurements of lesions on serial CT scans and defined PD as a ≥20% increase in the summed unidimensional measurements ([−30%, +20%]). These response criteria for unidimensional measurements were established for approximate consistency with volume changes that result when the WHO response criteria ([−50%, +25%]) are applied to bidimensional measurements of a spherical tumor.
The applicability of the RECIST guidelines to non-spherical tumors (such as retroperitoneal sarcoma or pleural mesothelioma) has been questioned. Spears [11] reported that diameter is an inaccurate estimate of area (and hence volume) for oblong tumors with length greater than three times width, and Mazumdar et al. [12] described greater disagreement between WHO and RECIST with increasingly elliptical tumors. Van Klaveren et al. [13] found similar evidence in 34 chemotherapy patients with pleural mesothelioma. Tumors were measured bidimensionally according to WHO and unidimensionally according to RECIST, with application of the respective tumor response criteria; poor agreement between the resulting response classifications was obtained despite the intended equivalence of RECIST and WHO.
Byrne and Nowak [14] recently published an alternative unidimensional measurement technique for mesothelioma that demonstrated strong correlation with bidimensional measurement in 73 chemotherapy patients. This “modified RECIST” technique dictates measurement of tumor thickness (i.e., short-axis dimension) instead of long-axis diameter and utilizes the same [−30%, +20%] response criteria as RECIST. The RECIST guidelines, however, might not provide response criteria that are appropriate for pleural mesothelioma, even when applied to a modified measurement technique. RECIST was developed based on mathematical models of spherical tumors, while mesothelioma presents a very different kind of growth. The RECIST guidelines present (1) a tumor measurement technique and (2) a set of tumor response criteria; as a modified measurement technique has been proposed, we suggest reconsideration of the response criteria as well.
We have generated mathematical models of mesothelioma tumors, which exhibit growth primarily in the short-axis dimension (thickness), to (1) calculate measurements of the tumor models based on different techniques and (2) equate those measurements with the volume changes that provide the foundation for tumor response classification. The purpose of this study was to demonstrate that the relationship between change in unidimensional measurements and change in volume is different in these mesothelioma tumor models than in spherical tumors. Furthermore, we have investigated the derivation of a distinct set of tumor response criteria for mesothelioma to accompany the new short-axis measurement technique.
Section snippets
Tumor models
Mesothelioma exhibits variable patterns of growth. This study modeled non-localized pleural thickening without fissure involvement, the classic growth pattern of malignant mesothelioma. Three geometric models were developed to represent the morphology of mesothelioma growth. The “annulus model” comprised a cylindrical shell, with growth occurring concentrically inward (Fig. 1A and B). The “lens model” comprised a cylindrical segment, with growth occurring perpendicular to the planar side of the
Assessment of response criteria
For each geometric tumor model, we analyzed the relationship between tumor growth and measurements of that growth (Fig. 2). Fig. 2A plots percent change in tumor volume versus percent change in measurement from baseline for a spherical tumor [9] and demonstrates the equivalence of different measurement response criteria for a sphere (the geometry that provided the foundation of RECIST). For example, a unidimensional measurement decrease of 30% (PR per RECIST) and a bidimensional measurement
Discussion
The “modified RECIST” measurement technique for mesothelioma measures tumor thickness instead of length; the RECIST response criteria then are applied to measurements obtained from this new technique. The results of the present study show that application of the RECIST response criteria ([−30%, +20%]) to measurements acquired with the modified RECIST technique for short-axis measurement of mesothelioma leads to response classification based on smaller volumetric tumor changes than intended by
Conclusion
The “modified RECIST” measurement technique for evaluating mesothelioma response to therapy recommends application of the standard RECIST tumor response criteria. We have demonstrated that the application of the RECIST response criteria to thickness measurements of mesothelioma models results in volumetrically different definitions of response for this tumor than for spherical tumors, which may result in over-classification of PR and PD. While our study does not propose definitive response
Acknowledgments
The authors would like to thank Heber MacMahon, M.D. and Joseph L. Ogarek, M.S. for assistance with the evaluation of clinical CT scans. Funded in part by The University of Chicago Cancer Research Center. S.A. holds warrants to shares of R2 Technology, Inc. (Sunnyvale, CA).
References (20)
Moving beyond chemotherapy: novel cytostatic agents for malignant mesothelioma
Lung Cancer
(2004)- et al.
A statistical simulation study finds discordance between WHO criteria and RECIST guidelines
J. Clin. Epidemiol.
(2004) - et al.
Inadequacy of the RECIST criteria for response evaluation in patients with malignant pleural mesothelioma
Lung Cancer
(2004) - et al.
Modified RECIST criteria for assessment of response in malignant pleural mesothelioma
Ann Oncol
(2004) - et al.
Evaluation of semi-automated measurements of mesothelioma tumor thickness on CT scans
Acad Radiol
(2005) - et al.
- et al.
Phase III study of pemetrexed in combination with cisplatin vs. cisplatin alone in patients with malignant pleural mesothelioma
JCO
(2003) - et al.
Cisplatin and gemcitabine treatment for malignant mesothelioma: a phase II study
JCO
(1999) - et al.
Phase II study of vinorelbine in patients with malignant pleural mesothelioma
JCO
(2000) Radiologic measurement of tumor size in clinical trials: past, present, and future
AJR
(2001)
Cited by (52)
Revised Modified Response Evaluation Criteria in Solid Tumors for Assessment of Response in Malignant Pleural Mesothelioma (Version 1.1)
2018, Journal of Thoracic OncologyMalignant pleural mesothelioma segmentation for photodynamic therapy planning
2018, Computerized Medical Imaging and GraphicsResponse evaluation in mesothelioma: Beyond RECIST
2015, Lung CancerCitation Excerpt :However, since MPM typically demonstrates a non-spherical growth pattern, it is often difficult to accurately categorize change in tumour burden using bi-dimensional or uni-dimensional tumour measurement [15]. One study evaluated the consistency of estimating disease volume by RECIST measurements [23] performed on geometric models. The study found that applying RECIST response criteria to a spherical tumour model resulted in partial response (PR) classification based on a 66% volume decrease and progressive disease (PD) based on a 73% volume increase.
Induction Therapy for Mesothelioma
2015, Seminars in Thoracic and Cardiovascular Surgery
- ☆
Presented in part at the International Association for the Study of Lung Cancer 11th World Conference on Lung Cancer, Barcelona, Spain, 2005.