Differences in target outline delineation from CT scans of brain tumours using different methods and different observers

doi:10.1016/S0167-8140(99)00015-8

Radiotherapy and Oncology

Volume 50, Issue 2, 1 February 1999, Pages 151-156

https://doi.org/10.1016/S0167-8140(99)00015-8 Get rights and content

Abstract

Purpose: To assess errors resulting from manual transfer of contour information for three-dimensional (3-D) target reconstruction, and to determine variations in target volume delineation of brain tumours by different radiation oncologists.

Materials and methods: Images of 18 patients with intracranial astrocytomas were used for retrospective treatment planning by five radiation oncologists. In this study, the target outline was delineated on sequential CT slices by an experienced radiation oncologist. Thereafter, the target outline was manually reconstructed by five radiation oncologists onto an A-P or lateral scout film. The same target outline was also reconstructed as a projection using the Beam's-eye view capability on a CT simulator unit. The two target outlines were compared by encompassing each shape with the smallest rectangle. The manually-reconstructed radiation field was termed ‘Field manually established on X-ray film (F-X)’, and the automatically-established field was termed ‘Field established by CT simulator (F-CT)’. In a second part of this study, four radiation oncologists defined contours from contrast enhanced CT images of nine patients with intracranial astrocytomas. The CT images of these nine cases included five pre-operative cases and four post-operative cases. Both gross tumour volume (GTV) and clinical target volume (CTV) were outlined on sequential CT slices. The target outlines for the four radiation oncologists were compared by identifying the smallest rectangular field surrounding the projection of these contours. The field established by each radiation oncologist was termed ‘Field of target volume (F-TV)’, and the overlapping portion of the four F-TVs for each case was termed ‘Overlapped field of the target volume (Fo-TV)’.

Results: The average distance between the isocentres of F-X and F-CT was 0.6±0.4 cm (mean±SD). The average ratio of the area of F-X divided by the area of F-CT was 1.04±0.12. The area of F-X was wider than the area of F-CT for four of the five oncologists. The ratio of the area of F-TV divided by the area of Fo-TV was calculated. The average ratio was relatively greater for CTV (2.07 in pre-operative cases and 2.11 in post-operative cases) than for GTV (1.12 in pre-operative cases and 1.41 in post-operative cases). Among radiation oncologists, variations in the delineation of GTV were smaller than those of CTV.

Conclusions: When using an X-ray simulator in treatment planning, errors resulting from the manual transfer of CT contour information to planar radiographs must be considered. When computer techniques are used to project contours onto radiographs errors resulting from individual variations when performing the contouring must be considered.

Introduction

Computed tomography (CT) is widely used in radiation oncology clinics. Fully integrated systems such as the CT simulator have also become widely available. Using such an integrated system, contours drawn on CT cross-sections can be accurately projected onto anterior-posterior or lateral images [9], [10]. However, conventional X-ray simulators are still used in many situations and the transfer of information from CT to the simulator radiograph is still a manual process. In this study, errors resulting from this transfer procedure are assessed.

The International Commission on Radiation Units and Measurements (ICRU) reported the definitions of target volumes, including gross tumour volume (GTV), clinical target volume (CTV), and planning target volume (PTV) [5]. Determining the target volume is one of the most important factors in treatment planning, but there is no standard method for making this determination. This study also examines variations in the definition of GTV and CTV for brain tumours contoured by different radiation oncologists.

Section snippets

The CT simulator system [9–11]

The original CT simulator at our institution was developed and was used clinically in 1987. In 1990, a second generation CT simulator with a new liquid crystal laser field projector was installed. In 1991, the network between the CT simulator and the Picture Archiving and Communication System (PACS) was developed. In 1993, a new CT scanner (CTS-20) and the latest computer were installed. A network connection between the multi-leaf collimator (MLC), the linear accelerator, and the planning

Accuracy of manual reconstruction (first study)

The average of 18 distances between the isocentre of F-X and that of F-CT for 18 cases reported by each of five radiation oncologists is shown in Table 1. That was 0.5±0.4 (mean±SD) cm, 0.6±0.3, 0.6±0.6, 0.6±0.2 and 0.8±0.3, respectively. The average of five radiation oncologists was 0.6±0.4 cm.

The average of 18 ratios of the area of F-X divided by that of F-CT for 18 cases reported by each of five radiation oncologists is also shown in Table 1. That was 1.05±0.16 (mean±SD), 1.03±0.11,

Discussion

The average distance between the isocentre (simply defined as the geometric centre) of the radiation field manually reconstructed by each radiation oncologist and the isocentre of the radiation field automatically established by the computer of the CT simulator on an A-P or lateral projection was about 0.6 cm. The accuracy of geometric reconstruction was limited within 1.0 cm in about 90%. The average ratio of the area of the manually reconstructed field divided by the area of the automatically

Acknowledgements

This work was presented at the 37th ASTRO annual meeting, Miami Beach, October 8–11, 1995. This work was partly supported by a grant-in-aid No. 08457243 and 09255225 of the Ministry of Education in Japan.

References (16)

B.W. Corn et al.
Conformal treatment of prostate cancer with improved targeting: superior prostate-specific antigen response compared to standard treatment
Int. J. Radiat. Oncol. Biol. Phys.
(1995)
G.A. Fortier et al.
MRI treatment planning for CNS neoplasm: graphic display of target volume using boxed cursor
Int. J. Radiat. Oncol. Biol. Phys.
(1990)
E. Glatstein et al.
The imaging revolution and radiation oncology: use of CT, ultrasound, and NMR for localization treatment planning and treatment delivery, Int
J. Radiat. Oncol. Biol. Phys.
(1985)
G. Leunens et al.
Quality assessment of medical decision making in radiation oncology: variability in target volume delineation for brain tumours
Radiother. Oncol.
(1993)
L.C. Myrianthopoulos et al.
Quantitation of treatment volumes from CT and MRI in high-grade gliomas: implications for radiotherapy
Magn. Reson. Imaging
(1992)
Y. Nagata et al.
CT simulator: a new 3-D planning and simulating system for radiotherapy: part 2
Clinical application, Int. J. Radiat. Oncol. Biol. Phys.
(1990)
Y. Nagata et al.
Three dimensional treatment planning for maxillary cancer using a CT simulator
Int. J. Radiat. Oncol. Biol. Phys.
(1994)
Y. Nagata et al.
Development of an integrated radiotherapy network system
Int. J. Radiat. Oncol. Biol. Phys.
(1996)

There are more references available in the full text version of this article.

Cited by (68)

Quality Assurance of Dose-Escalated Radiation Therapy in a Randomized Trial for Locally Advanced Oesophageal cancer
2019, International Journal of Radiation Oncology Biology Physics
Citation Excerpt :
Interestingly, we found that OAR delineations varied between participating centers because volumes were very heterogeneous. Inconsistencies in delineating target and OAR volumes have been identified for several other tumor sites33-36 and can compromise the validity of clinical trial results.37 This variability might be related to differences in opinion regarding target and OAR boundaries.
The ongoing phase 2/3 PRODIGE 26/CONCORDE trial compares chemoradiation therapy with and without dose escalation in patients with locally advanced or unresectable esophageal cancer. The results of a benchmark case procedure are reported here to evaluate the protocol compliance of participating centers as part of quality assurance for radiation therapy.
Volume delineation, target coverage, and dose constraints to the organs at risk (OARs) were assessed on treatment plans of a common benchmark case performed by each participating center. The centers were classified in 3 categories: per protocol, minor acceptable deviation (MiD), or major unacceptable deviation (MaD). A plan was rejected if ≥4 MiDs or 1 MaD were found.
Thirty-5 centers submitted 43 plans. Among them, 14 (32.6%) were per protocol, 19 (44.2%) presented at least 1 MiD, 2 (4.6%) presented at least 1 MaD, and 8 (18.6%) presented both MiD and MaD. Overall, 11 (25.6%) plans were rejected. Only 1 plan was rejected because gross tumor volume was not correctly delineated. The OAR delineation was respected in all cases. Dose constraints to the OARs were respected in the majority of cases except for the heart, where one-third of the plans presented a deviation. As for the target volume, 3 plans (5.8%) had a major underdosage and 1 plan (1.9%) had a major overdosage. Overall, 58% of all treatments were planned with intensity modulated radiation therapy, whereas 42% were planned with 3-dimensional chemoradiation therapy. Significantly more plans in the intensity modulated radiation therapy group were accepted compared with the 3-dimensional chemoradiation therapy group (P = .03).
The high frequency of protocol deviations underlines the importance of a quality assurance program in clinical trials. Further work should assess the impact of quality assurance for radiation therapy on patient outcomes.
Standardizing Normal Tissue Contouring for Radiation Therapy Treatment Planning: An ASTRO Consensus Paper
2019, Practical Radiation Oncology
Citation Excerpt :
According to ASTRO's publication, Safety Is No Accident, “[a] review of target delineation and image segmentation prior to planning deserves more standardization.”6 Multiple studies have shown variation in the delineation of radiation therapy targets and OARs in such disease sites as the central nervous system,7,8 head and neck,9,10 thorax,11-13 and pelvis.14,15 Differences in normal-tissue contouring can affect dosimetric plan optimization, dose volume histogram (DVH) results, and potentially normal tissue complication probabilities and clinical outcomes.
The comprehensive identification and delineation of organs at risk (OARs) are vital to the quality of radiation therapy treatment planning and the safety of treatment delivery. This guidance aims to improve the consistency of ontouring OARs in external beam radiation therapy treatment planning by providing a single standardized resource for information regarding specific OARs to be contoured for each disease site. The guidance is organized in table format as a quality assurance tool for practices and a training resource for residents and other radiation oncology students (see supplementary materials).
The Task Force formulated recommendations based on clinical practice and consensus. The draft manuscript was peer reviewed by 16 reviewers, the American Society for Radiation Oncology (ASTRO) legal counsel, and ASTRO's Multidisciplinary Quality Assurance Subcommittee and revised accordingly. The recommendations were posted on the ASTRO website for public comment in June 2018 for a 6-week period. The final document was approved by the ASTRO Board of Directors in August 2018.
Standardization improves patient safety, efficiency, and accuracy in radiation oncology treatment. This consensus guidance represents an ASTRO quality initiative to provide recommendations for the standardization of normal tissue contouring that is performed during external beam treatment planning for each anatomic treatment site. Table 1 defines 2 sets of structures for anatomic sites: Those that are recommended in all adult definitive cases and may assist with organ selection for palliative cases, and those that should be considered on a case-by-case basis depending on the specific clinical scenario. Table 2 outlines some of the resources available to define the parameters of general OAR tissue delineation.
Using this paper in conjunction with resources that define tissue parameters and published dose constraints will enable practices to develop a consistent approach to normal tissue evaluation and dose documentation.
CT- and MRI-based gross target volume comparison in vestibular schwannomas
2017, Reports of Practical Oncology and Radiotherapy
This study represents an enumeration and comparison of gross target volumes (GTV) as delineated independently on contrast-enhanced computed tomography (CT) and T1 and T2 weighted magnetic resonance imaging (MRI) in vestibular schwannomas (VS).
Multiple imaging in radiotherapy improves target localization.
42 patients of VS were considered for this prospective study with one patient showing bilateral tumor. The GTV was delineated separately on CT and MRI. Difference in volumes were estimated individually for all the 43 lesions and similarity was studied between CT and T1 and T2 weighted MRI.
The male to female ratio for VS was found to be 1:1.3. The tumor was right sided in 34.9% and left sided in 65.1%. Tumor volumes (TV) on CT image sets were ranging from 0.251 cc to 27.27 cc. The TV for CT, MRI T1 and T2 weighted were 5.15 ± 5.2 cc, 5.8 ± 6.23 cc, and 5.9 ± 6.13 cc, respectively. Compared to MRI, CT underestimated the volumes. The mean dice coefficient between CT versus T1 and CT versus T2 was estimated to be 68.85 ± 18.3 and 66.68 ± 20.3, respectively. The percentage of volume difference between CT and MRI (%VD: mean ± SD for T1; 28.84 ± 15.0, T2; 35.74 ± 16.3) and volume error (%VE: T1; 18.77 ± 10.1, T2; 23.17 ± 13.93) were found to be significant, taking the CT volumes as the baseline.
MRI with multiple sequences should be incorporated for tumor volume delineation and they provide a clear boundary between the tumor and normal tissue with critical structures nearby.
Practical implementation of failure mode and effects analysis for safety and efficiency in stereotactic radiosurgery
2015, International Journal of Radiation Oncology Biology Physics
To improve the safety and efficiency of a new stereotactic radiosurgery program with the application of failure mode and effects analysis (FMEA) performed by a multidisciplinary team of health care professionals.
Representatives included physicists, therapists, dosimetrists, oncologists, and administrators. A detailed process tree was created from an initial high-level process tree to facilitate the identification of possible failure modes. Group members were asked to determine failure modes that they considered to be the highest risk before scoring failure modes. Risk priority numbers (RPNs) were determined by each group member individually and then averaged.
A total of 99 failure modes were identified. The 5 failure modes with an RPN above 150 were further analyzed to attempt to reduce these RPNs. Only 1 of the initial items that the group presumed to be high-risk (magnetic resonance imaging laterality reversed) was ranked in these top 5 items. New process controls were put in place to reduce the severity, occurrence, and detectability scores for all of the top 5 failure modes.
FMEA is a valuable team activity that can assist in the creation or restructuring of a quality assurance program with the aim of improved safety, quality, and efficiency. Performing the FMEA helped group members to see how they fit into the bigger picture of the program, and it served to reduce biases and preconceived notions about which elements of the program were the riskiest.
Contouring variations and the role of atlas in non-small cell lung cancer radiation therapy: Analysis of a multi-institutional preclinical trial planning study
2015, Practical Radiation Oncology
Citation Excerpt :
The dose distribution of conformal radiation therapy plans is very dependent on the contours delineated on patient anatomy. Studies have shown variation in the contouring of targets and OARs in lung,2-6 head and neck,7,8 brain,9,10 prostate,11,12 and breast13,14 cases. The dosimetric impact of this contour variation can be significant,7,13,15 depending on the degree of variation and the plan dose gradient.
To quantify variations in target and normal structure contouring and evaluate dosimetric impact of these variations in non-small cell lung cancer (NSCLC) cases. To study whether providing an atlas can reduce potential variation.
Three NSCLC cases were distributed sequentially to multiple institutions for contouring and radiation therapy planning. No segmentation atlas was provided for the first 2 cases (Case 1 and Case 2). Contours were collected from submitted plans and consensus contour sets were generated. The volume variation among institution contours and the deviation of them from consensus contours were analyzed. The dose-volume histograms for individual institution plans were recalculated using consensus contours to quantify the dosimetric changes. An atlas containing targets and critical structures was constructed and was made available when the third case (Case 3) was distributed for planning. The contouring variability in the submitted plans of Case 3 was compared with that in first 2 cases.
Planning target volume (PTV) showed large variation among institutions. The PTV coverage in institutions’ plans decreased dramatically when reevaluated using the consensus PTV contour. The PTV contouring consistency did not show improvement with atlas use in Case 3. For normal structures, lung contours presented very good agreement, while the brachial plexus showed the largest variation. The consistency of esophagus and heart contouring improved significantly (t test; P < .05) in Case 3. Major factors contributing to the contouring variation were identified through a survey questionnaire.
The amount of contouring variations in NSCLC cases was presented. Its impact on dosimetric parameters can be significant. The segmentation atlas improved the contour agreement for esophagus and heart, but not for the PTV in this study. Quality assurance of contouring is essential for a successful multi-institutional clinical trial.
Segmentation algorithms of subcortical brain structures on MRI for radiotherapy and radiosurgery: A survey
2015, IRBM
This work covers the current state of the art with regard to approaches to segment subcortical brain structures. A huge range of diverse methods have been presented in the literature during the last decade to segment not only one or a constrained number of structures, but also a complete set of these subcortical regions. Special attention has been paid to atlas based segmentation methods, statistical models and deformable models for this purpose. More recently, the introduction of machine learning techniques, such as artificial neural networks or support vector machines, has helped the researchers to optimize the classification problem. These methods are presented in this work, and their advantages and drawbacks are further discussed. Although these methods have proved to perform well, their use is often limited to those situations where either there are no lesions in the brain or the presence of lesions does not highly vary the brain anatomy. Consequently, the development of segmentation algorithms that can deal with such lesions in the brain and still provide a good performance when segmenting subcortical structures is highly required in practice by some clinical applications, such as radiotherapy or radiosurgery.

View all citing articles on Scopus

View full text

Differences in target outline delineation from CT scans of brain tumours using different methods and different observers

Abstract

Introduction

Section snippets

The CT simulator system [9–11]

Accuracy of manual reconstruction (first study)

Discussion

Acknowledgements

Int. J. Radiat. Oncol. Biol. Phys.

Int. J. Radiat. Oncol. Biol. Phys.

J. Radiat. Oncol. Biol. Phys.

Radiother. Oncol.

Magn. Reson. Imaging

Clinical application, Int. J. Radiat. Oncol. Biol. Phys.

Int. J. Radiat. Oncol. Biol. Phys.

Int. J. Radiat. Oncol. Biol. Phys.