Usability of semiautomatic segmentation algorithms for tumor volume determination

Practical utility of liver segmentation methods in clinical surgeries and interventions

Clinical imaging (e.g., magnetic resonance imaging and computed tomography) is a crucial adjunct for clinicians, aiding in the diagnosis of diseases and planning of appropriate interventions. This is especially true in malignant conditions such as hepatocellular carcinoma (HCC), where image segmentation (such as accurate delineation of liver and tumor) is the preliminary step taken by the clinicians to optimize diagnosis, staging, and treatment planning and intervention (e.g., transplantation, surgical resection, radiotherapy, PVE, embolization, etc). Thus, segmentation methods could potentially impact the diagnosis and treatment outcomes. This paper comprehensively reviews the literature (during the year 2012–2021) for relevant segmentation methods and proposes a broad categorization based on their clinical utility (i.e., surgical and radiological interventions) in HCC. The categorization is based on the parameters such as precision, accuracy, and automation.

Multimodality 3D Tumor Segmentation in HCC Patients Treated with TACE

RATIONALE AND OBJECTIVES

To validate the concordance of a semiautomated multimodality lesion segmentation technique between contrast-enhanced magnetic resonance imaging (CE-MRI), cone-beam computed tomography (CBCT), and multidetector CT (MDCT) in patients with hepatocellular carcinoma (HCC) treated with transarterial chemoembolization (TACE).

MATERIALS AND METHODS

This retrospective analysis included 45 patients with unresectable HCC who underwent baseline CE-MRI within 1 month before the treatment, intraprocedural CBCT during conventional TACE, and MDCT within 24 hours after TACE. Fourteen patients were excluded because of atypical lesion morphology, portal vein invasion, or small lesion size which precluded sufficient lesion visualization. Thirty-one patients with a total of 40 target lesions were included into the analysis. A tumor segmentation software, based on non-Euclidean geometry and theory of radial basis functions, was used to allow for the segmentation of target lesions in 3D on all three modalities. The algorithm created image-based masks located in a 3D region whose center and size were defined by the user, yielding the nomenclature "semiautomatic". On the basis of that, tumor volumes on all three modalities were calculated and compared using a linear regression model (R(2) values). Residual plots were used to analyze drift and variance of the values.

RESULTS

The mean value of tumor volumes was 18.72 ± 19.13 cm(3) (range, 0.41-59.16 cm(3)) on CE-MRI, 21.26 ± 21.99 cm(3) (range, 0.62-86.82 cm(3)) on CBCT, and 19.88 ± 20.88 cm(3) (range, 0.45-75.24 cm(3)) on MDCT. The average volumes of the tumor were not significantly different between CE-MRI and DP-CBCT, DP-CBCT and MDCT, MDCT and CE-MRI (P = .577, .770, and .794, respectively). A strong correlation between volumes on CE-MRI and CBCT, CBCT and MDCT, MDCT and CE-MRI was observed (R(2) = 0.974, 0.992 and 0.983, respectively). When plotting the residuals, no drift was observed for all methods showing deviations of no >10% of absolute volumes (in cm(3)).

CONCLUSIONS

A semiautomated 3D segmentation of HCC lesions treated with TACE provides high volumetric concordance across all tested imaging modalities.

Comparison of semi-automatic volumetric VX2 hepatic tumor segmentation from cone beam CT and multi-detector CT with histology in rabbit models

RATIONALE AND OBJECTIVES

The purpose of this study was to compare tumor volume in a VX2 rabbit model as calculated using semiautomatic tumor segmentation from C-arm cone-beam computed tomography (CBCT) and multidetector computed tomography (MDCT) to the actual tumor volume.

MATERIALS AND METHODS

Twenty VX2 tumors in 20 adult male New Zealand rabbits (one tumor per rabbit) were imaged with CBCT (using an intra-arterial contrast medium injection) and MDCT (using an intravenous contrast injection). All tumor volumes were measured using semiautomatic three-dimensional volumetric segmentation software. The software uses a region-growing method using non-Euclidean radial basis functions. After imaging, the tumors were excised for pathologic volume measurement. The imaging-based tumor volume measurements were compared to the pathologic volumes using linear regression, with Pearson's test, and correlated using Bland-Altman analysis.

RESULTS

Average tumor volumes were 3.5 ± 1.6 cm(3) (range, 1.4-7.2 cm(3)) on pathology, 3.8 ± 1.6 cm(3) (range, 1.3-7.3 cm(3)) on CBCT, and 3.9 ± 1.6 (range, 1.8-7.5 cm(3)) on MDCT (P < .001). A strong correlation between volumes on pathology and CBCT and also with MDCT was observed (Pearson's correlation coefficient = 0.993 and 0.996, P < .001, for CBCT and MDCT, respectively). Bland-Altman analysis showed that MDCT tended to overestimate tumor volume, and there was stronger agreement for tumor volume between CBCT and pathology than with MDCT, possibly because of the intra-arterial contrast injection.

CONCLUSIONS

Tumor volume as measured using semiautomatic tumor segmentation software showed a strong correlation with the "real volume" measured on pathology. The segmentation software on CBCT and MDCT can be a useful tool for volumetric hepatic tumor assessment.

Usefulness of semi-automatic volumetry compared to established linear measurements in predicting lymph node metastases in MSCT

BACKGROUND

Volumetry of lymph nodes potentially better reflect asymmetric size alterations independently of lymph node orientation in comparison to metric parameters (e.g. long-axis diameter).

PURPOSE

To distinguish between benign and malignant lymph nodes by comparing 2D and semi-automatic 3D measurements in MSCT.

MATERIAL AND METHODS

FDG-18 PET-CT was performed in 33 patients prior to therapy for malignant melanoma at stage III/IV. One hundred and eighty-six cervico-axillary, abdominal and inguinal lymph nodes were evaluated independently by two radiologists, both manually and with the use of semi-automatic segmentation software. Long axis (LAD), short axis (SAD), maximal 3D diameter, volume and elongation were obtained. PET-CT, PET-CT follow-up and/or histology served as a combined reference standard. Statistics encompassed intra-class correlation coefficients and ROC curves.

RESULTS

Compared to manual assessment, semi-automatic inter-observer variability was found to be lower, e.g. at 2.4% (95% CI 0.05-4.8) for LAD. The standard of reference revealed metastases in 90 (48%) of 186 lymph nodes. Semi-automatic prediction of lymph node metastases revealed highest areas under the ROC curves for volume (reader 1 0.77, 95%CI 0.64-0.90; reader 2 0.76, 95%CI 0.59-0.86) and SAD (reader 1 0.76, 95%CI 0.64-0.88; reader 2 0.75, 95%CI 0.62-0.89). The findings for LAD (reader 1 0.73, 95%CI 0.60-0.86; reader 2 0.71, 95%CI 0.71, 95%CI 0.57-0.85) and maximal 3D diameter (reader 1 0.70, 95%CI 0.53-0.86; reader 2 0.76, 95%CI 0.50-0.80) were found substantially lower and for elongation (reader 1 0.65, 95%CI 0.50-0.79; reader 2 0.66, 95%CI 0.52-0.81) significantly lower (p < 0.05).

CONCLUSION

Semi-automatic analysis of lymph nodes in malignant melanoma is supported by high segmentation quality and reproducibility. As compared to established SAD, semi-automatic lymph node volumetry does not have an additive role for categorizing lymph nodes as normal or metastatic in malignant melanoma.

Liver volume measurement: reason of the difference between in vivo CT-volumetry and intraoperative ex vivo determination and how to cope it

Purpose

Volumetric assessment of the liver regularly yields discrepant results between pre- and intraoperatively determined volumes. Nevertheless, the main factor responsible for this discrepancy remains still unclear. The aim of this study was to systematically determine the difference between in vivo CT-volumetry and ex vivo volumetry in a pig animal model.

Material and Methods

Eleven pigs were studied. Liver density assessment, CT-volumetry and water displacement volumetry was performed after surgical removal of the complete liver. Known possible errors of volume determination like resection or segmentation borders were eliminated in this model. Regression analysis was performed and differences between CT-volumetry and water displacement determined.

Results

Median liver density was 1.07 g/ml. Regression analysis showed a high correlation of r² = 0.985 between CT-volumetry and water displacement. CTvolumetry was found to be 13% higher than water displacement volumetry (p < 0.0001).

Conclusion

In this study the only relevant factor leading to the difference between in vivo CT-volumetry and ex vivo water displacement volumetry seems to be blood perfusion of the liver. The systematic difference of 13 percent has to be taken in account when dealing with those measures.

Hepatocellular carcinoma: CT for tumor response after transarterial chemoembolization in patients exceeding Milan criteria--selection parameter for liver transplantation

PURPOSE

To retrospectively evaluate the clinical outcome of patients with hepatocellular carcinoma (HCC) who exceeded the Milan criteria, who underwent transarterial chemoembolization (TACE) before orthotopic liver transplantation (OLT), to determine the value of computed tomography (CT)-based tumor response to TACE as a preoperative selection criterion for OLT.

MATERIALS AND METHODS

The study included 33 patients with HCC who exceeded the Milan criteria and underwent OLT after TACE. Informed written consent was obtained before TACE and OLT. Institutional review board approval was not required. Tumor response to TACE was evaluated at 1 month with CT according to amended Response Evaluation Criteria in Solid Tumours (RECIST) guidelines. In the explanted liver, degree of tumor necrosis (> or = 90%, 50%-89%, or < 50%), residual tumor stage and grade, and presence of microvascular invasion were assessed. Follow-up after OLT ranged from 1 to 143 months.

RESULTS

After TACE, CT showed complete tumor response (CR) in 18 (55%) patients. On the explanted liver, tumor necrosis was rated 90% or greater in 20 (61%) patients, with a good correlation with CT. Microvascular invasion was observed in nine (27%) of 33 patients; none of them were reported to have a CR at CT. The 5-year cumulative survival rate after OLT was 72.5%; it was significantly (P = .003) higher in patients with a CR (94.4%) compared with patients with a partial response (PR) (45.4%) and stable disease (50%). The 5-year cumulative recurrence-free rate after OLT was 74.4%; it was not affected by the tumor nodule size and number, whereas it was significantly (P = .008) higher in patients with a CR (94.4%) compared with patients with a PR (46.7%) and stable disease (50%).

CONCLUSION

In patients with HCC who exceeded the Milan criteria, a CR after TACE, on the basis of amended RECIST guidelines, is associated with excellent posttransplantation outcomes. Therefore, 1-month response to TACE assessed at CT may represent a valid selection criterion for OLT.

Liver tumour segmentation using contrast-enhanced multi-detector CT data: performance benchmarking of three semiautomated methods

OBJECTIVE

Automatic tumour segmentation and volumetry is useful in cancer staging and treatment outcome assessment. This paper presents a performance benchmarking study on liver tumour segmentation for three semiautomatic algorithms: 2D region growing with knowledge-based constraints (A1), 2D voxel classification with propagational learning (A2) and Bayesian rule-based 3D region growing (A3).

METHODS

CT data from 30 patients were studied, and 47 liver tumours were isolated and manually segmented by experts to obtain the reference standard. Four datasets with ten tumours were used for algorithm training and the remaining 37 tumours for testing. Three evaluation metrics, relative absolute volume difference (RAVD), volumetric overlap error (VOE) and average symmetric surface distance (ASSD), were computed based on computerised and reference segmentations.

RESULTS

A1, A2 and A3 obtained mean/median RAVD scores of 17.93/10.53%, 17.92/9.61% and 34.74/28.75%, mean/median VOEs of 30.47/26.79%, 25.70/22.64% and 39.95/38.54%, and mean/median ASSDs of 2.05/1.41 mm, 1.57/1.15 mm and 4.12/3.41 mm, respectively. For each metric, we obtained significantly lower values of A1 and A2 than A3 (P < 0.01), suggesting that A1 and A2 outperformed A3.

CONCLUSIONS

Compared with the reference standard, the overall performance of A1 and A2 is promising. Further development and validation is necessary before reliable tumour segmentation and volumetry can be widely used clinically.

Morphological analysis of pancreatic adenocarcinoma on multidetector row computed tomography: implications for treatment response evaluation

OBJECTIVES

Response Evaluation Criteria in Solid Tumors (RECIST) guidelines assume spherical shape of tumors. Morphology of pancreatic adenocarcinoma (PAC) on multidetector row computed tomography was investigated to evaluate the applicability of RECIST guidelines.

METHODS

Study population comprised 16 patients with histologically confirmed localized PAC enrolled in a phase II clinical trial of chemoradiation. Pancreatic adenocarcinomas were segmented on baseline and follow-up multidetector row computed tomography with commercially available software. Tumor volumes (mL), RECIST diameter (mm), volume equivalent sphere diameter (VESD, mm), maximum 3-dimensional diameter (M3DD, mm), and elongation value were obtained. RECIST diameter, VESD and M3DD of the tumors at baseline and follow-up were compared to determine differences. Elongation values were analyzed. The significance level was set at P less than 0.05.

RESULTS

Mean volume, RECIST diameter, VESD, M3DD, and elongation for baseline versus follow-up studies were 23.12 mL versus 19.43 mL (P > 0.05), 41.86 mm versus 39.35 mm (P > 0.05), 33.14 mm versus 32.1 mm (P > 0.05), 51.76 mm versus 51.73 mm (P > 0.05), and 0.67 versus 0.76 (P > 0.05), respectively. There was a significant difference at baseline and follow-up between RECIST diameter, VESD, and M3DD (P < 0.05, in all instances).

CONCLUSIONS

Our results suggest that PACs are not spherical in shape. Evaluation of PAC treatment response based on RECIST guidelines may not be accurate.

[Volumetry of metastases from renal cell carcinoma: comparison with the RECIST criteria]

For patients with metastasized renal cell carcinoma (RCC), imaging techniques are of great importance. Currently, therapy widely relies on antiangiogenic factors, which frequently lead to relatively subtle changes in the size of lesions. From this aspect the commonly used RECIST criteria (response evaluation criteria in solid tumors) must be considered as imprecise for the evaluation of the response to therapy. This article gives a review on new software-based volumetric methods, which allow therapy-induced changes in the size of metastases from RCC to be detected with higher sensitivity and reproducibility. A comparison of RECIST and volumetry was carried out with data from patients with metastasized RCC to demonstrate the higher sensitivity of the 3D volumetric procedure.

Early clinical evaluation of a novel three-dimensional structure delineation software tool (SCULPTER) for radiotherapy treatment planning

Modern radiotherapy treatment planning (RTP) necessitates increased delineation of target volumes and organs at risk. Conventional manual delineation is a laborious, time-consuming and subjective process. It is prone to inconsistency and variability, but has the potential to be improved using automated segmentation algorithms. We carried out a pilot clinical evaluation of SCULPTER (Structure Creation Using Limited Point Topology Evidence in Radiotherapy) - a novel prototype software tool designed to improve structure delineation for RTP. Anonymized MR and CT image datasets from patients who underwent radiotherapy for bladder or prostate cancer were studied. An experienced radiation oncologist used manual and SCULPTER-assisted methods to create clinically acceptable organ delineations. SCULPTER was also tested by four other RTP professionals. Resulting contours were compared by qualitative inspection and quantitatively by using the volumes of the structures delineated and the time taken for completion. The SCULPTER tool was easy to apply to both MR and CT images and diverse anatomical sites. SCULPTER delineations closely reproduced manual contours with no significant volume differences detected, but SCULPTER delineations were significantly quicker (p<0.05) in most cases. In conclusion, clinical application of SCULPTER resulted in rapid and simple organ delineations with equivalent accuracy to manual methods, demonstrating proof-of-principle of the SCULPTER system and supporting its potential utility in RTP.

Target Definition in Prostate, Head, and Neck

Target definition is a major source of errors in both prostate and head and neck external-beam radiation treatment. Delineation errors remain constant during the course of radiation and therefore have a large impact on the dose to the tumor. Major sources of delineation variation are visibility of the target including its extensions, disagreement on the target extension, and interpretation or lack of delineation protocols. The visibility of the target can be greatly improved with the use of multimodality imaging. Both in the head and neck and the prostate, computed tomography (CT)-magnetic resonance imaging coregistration decreases the target volume and its variability. CT-positron emission tomography delineation is promising for delineation in head and neck cancer. Despite the better visibility, a different interpretation of the target extension remains a major source of error. The use of coregistration of CT with a second modality, together with improved guidelines for delineation and an online anatomical atlas, increases agreement between observers in prostate, lung, and nasopharynx tumors. Delineation errors should not be treated differently from other geometrical errors. Similar margin recipes for the correction of setup errors and organ motion should be adapted to incorporate the effect of delineation errors. A calculation of a 3-dimensional clinical target volume-planning target volume margin incorporating delineation errors for the head and neck is around 6.1 to 9.7 mm. Given the good local control of IMRT with smaller margins and smaller pathological specimens, it is likely that the delineated CTV frequently overestimates the actual volume.

[Computer-assisted operational planning for pediatric abdominal surgery. 3D-visualized MRI with volume rendering]

Exact surgical planning is necessary for complex operations of pathological changes in anatomical structures of the pediatric abdomen. 3D visualization and computer-assisted operational planning based on CT data are being increasingly used for difficult operations in adults. To minimize radiation exposure and for better soft tissue contrast, sonography and MRI are the preferred diagnostic methods in pediatric patients. Because of manifold difficulties 3D visualization of these MRI data has not been realized so far, even though the field of embryonal malformations and tumors could benefit from this.A newly developed and modified raycasting-based powerful 3D volume rendering software (VG Studio Max 1.2) for the planning of pediatric abdominal surgery is presented. With the help of specifically developed algorithms, a useful surgical planning system is demonstrated. Thanks to the easy handling and high-quality visualization with enormous gain of information, the presented system is now an established part of routine surgical planning.