| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Fondazione Biomedica Europea–onlus, Via Nizza, 53–00198, Rome,
Italy.
2 Department of Radiology, University of Rotterdam 30156D, The
Netherlands.
3 Department of Surgery, University of Parma 43100, Italy.
4 Department of Radiology, University of Parma 43100, Italy.
Received October 20, 2003;
accepted after revision May 28, 2004.
Address correspondence to G. Luccichenti.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Fifteen patients with rectosigmoid cancer underwent helical CT before and after neoadjuvant radiation therapy. The helical CT examination was performed after colon distention with air and IV administration of an antiperistaltic drug. Two scans were obtained: one with the patient in the supine position and the other with the patient in the prone position after contrast medium injection. The maximal wall thickness and the volumetric analysis of the tumor were obtained through manual segmentation.
RESULTS. The mean of the differences between the volumetric analysis of the scans obtained before and after radiation therapy was 8.3 ± 10.3 (SD) mL (–22.7%) (p < 0.05). The mean of the differences between the maximal wall thickness of the pre– and post–radiation therapy scans was 3.4 ± 2.6 mm (–19.1%) (p < 0.05). A significant difference was observed between the variation of the maximal wall thickness and the variation of volumetric analysis in pre– and post–radiation therapy scans (p < 0.05). The patients could be classified in different response categories depending on the measurement method and on the response criteria.
CONCLUSION. Volumetric analysis of rectosigmoid cancer is feasible. A long-term study is needed to correlate volumetric assessment with patient outcome.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
In radiologic practice, tumor volume is assumed to be representative of tumor cellularity [6, 7]. No reliable methods exist for volume estimation of rectosigmoid cancer. World Health Organization (WHO) [6] and Response Evaluation Criteria in Solid Tumors (RECIST) guidelines [8] do not provide an indication for neoplasms involving structures with complex cross-sectional anatomy or hollow viscera. Therefore, the assessment of tumor response to neoadjuvant therapy is based mainly on the subjective judgment of the radiologist. Subjective volume estimation provides poor results [9]. Volumetric analysis enables quantitative volumetric assessment of organs and lesions and provides reliable and more reproducible data [10, 11].
In this study, conventional linear measurements and volumetric analysis were compared in the assessment of patients with rectosigmoid cancer before and after radiation therapy, applying WHO and RECIST guidelines.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The patients underwent radiation therapy. Irradiation was applied with one mid dorsal and two angled dorsal radiation portals with the patient in the prone position to include the rectum and the surrounding tissues. A total dose of 4,000 cGy (250 cG daily) was given 4 days per week for 4 weeks. CT was performed within 1 week before and 1 week after radiation therapy.
We used a protocol for helical CT that is similar to the ones applied for virtual colonography. Colonic cleansing was achieved with a standard bowel preparation (4 L of polyethylene glycol solution the day before CT). After colon distention with air and IV administration of an antiperistaltic drug (10 mg of butyl scopolamine), helical CT of the abdomen was performed using a single-detector scanner (Somatom Plus 4, Siemens Medical Solutions). Insufflation was performed with the patient in the supine position using a 2-French soft latex catheter previously inserted into the rectum. Insufflation was complete when the patient's tolerance was reached. The catheter was not removed, and reinsufflation was performed if the distention was considered inadequate on the topogram obtained before each scan. Patients were scanned using the following parameters: collimation, 3 mm; table feed, 6 mm per rotation; reconstruction interval, 1 mm. One hundred milliliters of IV contrast material (300 mg I/mL Iomeron [iomeprol], Bracco Diagnostics) was administered through an antecubital vein at a rate of 3 mL/sec for each scan. Two scans were obtained after contrast medium injection: one with the patient supine and the other with the patient prone. Patients were instructed to hold their breath during scanning. Scans were reconstructed with a soft-tissue convolution kernel, and images were sent to a workstation equipped with 3D reconstruction software (Vitrea 2.2, Vital Images).
For each study, a radiologist experienced in gastrointestinal cross-sectional imaging determined the scan position (supine or prone) in which the best distention of rectum and sigmoid was present. The measurements of the tumor were then performed using this scan position on the pre– and post–radiation therapy CT examinations.
Two radiologists, experienced in gastrointestinal cross-sectional imaging, performed the measurements blinded from the results of each other and from the results of conventional colonoscopy. The assessment of tumor response was performed by measuring the maximal wall thickness on axial slices at the same level in the pre– and post–radiation therapy scans.
The assessment of tumor response using volumetric analysis was performed through manual segmentation by defining the tumor boundaries slice-byslice (Figs. 1A, 1B, 1C and 2A, 2B). A hypodense structure without contrast material enhancement with gas bubbles or one with gas bubbles was considered suspicious for stool remnants. Differentiation of fibrosis and reactive inflammation from residual tumor in the perirectal fat was mainly based on contrast enhancement and lesion conspicuity. Marks were drawn at the proximal and distal limits of the tumor. When the neoplasm involved the whole circumference, the endoluminal content was also included. In this case, a second manual segmentation including the lumen alone was performed between the two marks. Tumor volume was then obtained by subtracting the volume of the lumen from the entire volume. The mean of the measurements of the two observers was used for all parameters.
|
|
|
|
|
We assessed tumor response to radiation therapy using the WHO and RECIST criteria to characterize changes in maximal wall thickness. Complete or partial response and stable or progressive disease are defined in terms of variation in the percentage of tumor volume. With the WHO criteria, partial or complete response is defined by a reduction of the maximal wall thickness of greater than 50% and 100% (i.e., no measurable disease), respectively [6]. With the RECIST criteria, partial or complete response is defined by a reduction of the maximal wall thickness of greater than 30% and 100% (i.e., no measurable disease), respectively [8].
At present, there are no criteria by which to assess response to treatment using the volumetric analysis technique; therefore, for this study, we established criteria to mirror the concepts of the WHO and RECIST guidelines for application to a volumetric analysis technique. A 30% variation in the largest diameter of a lesion (RECIST criteria for partial response) corresponds to a variation of almost 50% of the square (WHO criteria for partial response) and to a variation of 65.7% of the cube of this largest diameter (Fig. 3). To better illustrate the rationale behind the threshold for partial response, let us consider a lesion with a largest diameter of 30 mm: A reduction of the largest diameter from 30 to 21 mm corresponds to a 30% variation, which is the threshold for partial response according to the RECIST criteria. With the WHO criteria, variation is estimated using two perpendicular diameters: If both are 30 mm, the product of these diameters is 900 mm2. Hence, a 30% variation of these diameters from 30 to 21 mm leads to a variation of the product from 900 to 441 mm2, which is almost a 50% (49%) variation. With the same procedure, variation can be estimated in the three dimensions: If the three diameters are 30 mm, the variation of all these diameters from 30 to 21 mm leads to a variation of the product of these diameters from 27,000 to 9,261 mm3, which is a variation of 65.7%. We applied this latter percentage as threshold for partial response using a volumetric analysis technique.
|
Statistical analysis was performed using the Student's t test for paired data and Pearson's correlation test. A p value less than 0.05 was considered significant.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The mean maximal wall thickness measured on scans was 18.0 ± 7.4 mm before radiation therapy, and it was 14.6 ± 7.4 mm after radiation therapy. The mean of the difference between the maximal wall thicknesses shown on pre– and post–radiation therapy scans was 3.4 ± 2.6 mm. The mean variation in maximal wall thickness between the pre– and post–radiation therapy scans expressed as a percentage was –19.1%. A significant difference was observed between maximal wall thicknesses on pre– and post–radiation therapy scans (p < 0.05).
The mean of the differences between the volumetric analysis values of the pre– and post–radiation therapy scans was 8,313 ± 10,258 mm3. The mean variation in volumetric analysis results between pre– and post–radiation therapy scans expressed as a percentage was –22.7%. A significant difference was observed between the volumetric analysis on pre– and post–radiation therapy scans (p < 0.05).
A significant difference was observed between the variation in maximal wall thickness and the variation in the volumetric analysis value on pre– and post–radiation therapy scans (p < 0.05). The correlation between the variation in maximal wall thickness and the variation in volume on pre– and post–radiation therapy was not significant (r = 0.44, p > 0.05).
The variation in the maximal wall thickness was also compared with the variation in the cubic root of volumetric analysis value on the pre– and post–radiation therapy scans. The cubic root of the volumetric analysis value was calculated to make volumetric analysis of the same dimension of maximal wall thickness, regardless of the shape of the lesion. In other words, the volumetric analysis values expressed in centimeters cubed were converted into centimeters to allow comparison with the maximal wall thickness values. A significant difference was observed between the variation in maximal wall thickness and the cubic root of the variation of volumetric analysis on pre– and post–radiation therapy scans (p < 0.05). The correlation between the variation in the maximal wall thickness and the cubic root of the variation in volumetric analysis values on pre– and post–radiation therapy scans was not significant (r = 0.34, p > 0.05).
According to the WHO guidelines, one patient had a partial response to therapy and the remaining 14 patients were stable. According to the RECIST guidelines, four patients had partial response, and the remaining were stable. When we used the 65.7% threshold for variation in the volumetric analysis values as the threshold for partial response, all the patients were stable.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The variation in tumor bulk is significantly different for the maximal wall thickness and volumetric analysis techniques. This difference is not only due to the diverse scales (maximal wall thickness is measured in millimeters; volumetric analysis, in millimeters cubed), because significant differences between maximal wall thickness and the cubic root of the volumetric analysis values were observed. These results can be explained by the fact that the WHO and RECIST guidelines are heavily dependent on mathematic [15] and geometric [7, 9, 11, 15, 16] issues, which are more prominent in the case of irregular and hollow structures. The lack of correlation between tumor volume, which is a 3D parameter, and 2D or one-dimensional measurements is compensated for using threshold criteria with ordinal scales based on percentage variation. However, ordinal scales allow classification rather than quantification of tumor response.
Concerning the geometric issues, linear measurements can hardly be applied to estimate the volume of irregular lesions. The tumor may vary in one dimension, but not in the maximal wall thickness. The linear measurement is affected by the anatomy of the viscera. Volumetric analysis techniques can reduce the effect of these issues because the volume is calculated directly by counting the number of voxels forming the lesion. CT phantom studies have already shown that the accuracy of volumetric measurements techniques is higher than that of single- or bidimensional techniques [9]. The true lesion volume of a pathologic specimen cannot be measured reliably because the spatial relationships are lost and the intra- and extracellular fluids and blood content are altered [17]. Moreover, for oncologic follow-up, the relevant information is the variation in tumor bulk rather than the absolute measurement.
In this study, we set the threshold for volumetric analysis at 65.7% and found that response to radiation therapy was underestimated when this threshold was used compared with when conventional WHO and RECIST guidelines were applied. The cases could be reclassified in different response categories because of the different criteria used for the guidelines. In other studies, researchers have reported that 17.7–26% of cases can be reclassified depending on the measurement method and on the response criteria [7, 16, 18]. Even though a gold standard is not available for validation, this problem can be overcome with a long-term survival study in which the volumetric analysis is correlated with patient outcome.
A potential source of variability in the volumetric analysis technique is related to the expandable and flexible nature of the rectosigmoid colon. The degree of distention affects the volume of the organ and the wall thickness [19]. Poor distention prevents the proper assessment of the location and length of wall thickening [20]. In this study, the colon was reinflated if partial collapses were observed at the level of rectosigmoid junction on the topogram before each scan was obtained.
Segmentation of the tumor should be performed to avoid categorizing intraluminal residues and reactive fibrosis as "tumor." The amount of residual fluid and fecal material depends on the bowel preparation. Fecal material can be differentiated from the lesion by the presence of air because lesions remain on the dependent surface and because of the lack of contrast material enhancement [21]. A lesion is a mixture of neoplastic cells, vessels, connective tissue, interstitium, necrosis, and fluid collections. An increase in the volume of one of the noncellular components (i.e., necrosis) may mask a decrease in cell number. Conventional CT and MR scans do not provide any information about the viability of the tumor, which can be assessed on PET. The combined use of morphologic and functional measurements could increase the accuracy of the quantification of tumor response to therapy.
This study is the first attempt to combine volumetric analysis with the virtual colonography technique for oncologic follow-up of hollow viscera. The technique results were feasible and might be an additional tool for the assessment of neoadjuvant radiation therapy. Volumetric analysis might also allow accurate quantification of tumor volume rather than allowing only classification of patients into response groups.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  B. Barbaro, C. Fiorucci, C. Tebala, V. Valentini, M. A. Gambacorta, F. M. Vecchio, G. Rizzo, C. Coco, A. Crucitti, C. Ratto, et al. Locally Advanced Rectal Cancer: MR Imaging in Prediction of Response after Preoperative Chemotherapy and Radiation Therapy Radiology, March 1, 2009; 250(3): 730 - 739. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. J. Pickhardt, V. T. Lehman, T. C. Winter, and A. J. Taylor Polyp volume versus linear size measurements at CT colonography: implications for noninvasive surveillance of unresected colorectal lesions. Am. J. Roentgenol., June 1, 2006; 186(6): 1605 - 1610. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
