Software development for hepatopulmonary shunt estimation by gamma camera method in transarterial radioembolization

Abstract

Trans-arterial radioembolization (TARE) is an established treatment for inoperable hepatocellular carcinoma and liver metastases from Carcinoma of gastrointestinal tract. Radiation-induced toxicity to the lung parenchyma is the dose-limiting factor in TARE. Pretreatment hepatopulmonary shunt (HPS) is estimated by gamma camera method by transarterial administration of 370MBq ^99mTc macro aggregated albumin. We have developed HPS software on XELERIS-1.123 workstation, GE medical systems, Milwaukee, USA, for accurate calculation of HPS. This software has also been tested on a higher version of XELERIS workstation, and it has been found to work well in all versions.

Keywords

Hepatopulmonary shunt - transarterial radioembolization - Y-90 microspheres

Introduction

Transarterial radioembolization (TARE) using Y-90 labeled microspheres is an established treatment modality for primary as well as secondary hepatic malignant tumor.[1],[2],[3],[4],[5] There are two Food and Drug Authority, USA, approved commercial90 Y-microsphere products are available in the market, i.e.,90 Y-SIRSphere by SIRTEX, Sydney, Australia, and90 Y-TheraSphere by MDS Nordion, Ottawa, Canada.[6],[7] TARE requires multidisciplinary therapeutic approach, to circumvent the therapy-related complications. To minimize the TARE-related complications, mapping of microsphere distribution becomes very important along with super selective delivery of microsphere.99m Tc macro aggregated albumin (99m Tc-MAA) is used as surrogate to map the post-therapeutic microsphere distribution in hepatic tumor, normal liver, shunt in lung, and any other abnormal distribution such as in stomach, duodenum, and gallbladder. Usually, for planning the treatment, a planar and single-photon emission computed tomography (SPECT) studies are performed on gamma camera after transarterial infusion of99m Tc-MAA in intervention radiology.[8],[9],[10] Extra hepatic uptake of tracer can be seen either in stomach, gallbladder, and duodenum due to collateral arterial supply from hepatic artery or in lung due to arteriovenous shunt in tumour resulting to hepatopulmonary shunt (HPS).[11] Uptake in stomach and duodenum can be prevented by super selective delivery of microsphere and coiling the collateral artery which supplies to these organs.[6],[7],[8] HPS results in radiation dose to the lung and there is no way either to reduce the unwarranted exposure or prevent it. Hence, HPS is a limiting factor in therapeutic dose calculation in TARE. A radiation dose of more than 30 Gy in single setting and 50 Gy in a life time to the lung is a contraindication to carry out the therapeutic procedure because it causes irreversible radiation-induced injury to the lung. More than 20% HPS value requires dose modification and above 30% HPS, TARE cannot be performed.[12],[13],[14],[15] Hence, accurate evaluation of HPS is of utmost importance to avoid complications.[9],[10] HPS evaluation is performed using planar or SPECT study. We have developed software for HPS estimation using planar gamma camera imaging and evaluated it.

Materials and Methods

Software development

HPS Software has been developed using Aladdin editor on Xeleris-1.1234 workstation, GE medical systems, Milwaukee, USA.

Software design

This Software has two main divisions, left panel [Figure 1A] and right container [Figure 1B]. The left panel contains two buttons-Main button and HPS button [Figure 1A]1, and three panels – color panel, zoom panel, and text panel [Figure 1A]2, right container is divided into two columns and three rows making a total of six image cells. Upper two rows (i.e., 4 image cells) of image container are used to load images, and third row is used to display results. In first row anterior and posterior images of thorax including lung and liver of patient with Co-57 phantom (SET-1) are loaded [Figure 1B]1 and [Figure 1]1B2 and in second row anterior and posterior images without phantom (SET-2) of the same area are loaded [Figure 1B]3 and [Figure 1]1B4 and in third row results are displayed in [Figure 1B]5 and [Figure 1B]6. Software algorithm is shown in [Figure 2].

Figure 1 This is the display of the hepatopulmonary shunt software has left (A) and right (B) devision; (A1): Contains main button and hepatopulmonary shunt button. (A2): Contains zoom panel, color panel, and text panel. (B1 and B2): Contains SET-1 image in anterior and posterior view. (B3 and B4): Contains SET-2 image in anterior and posterior view. (B5 and B6): Displays background corrected region of interest counts, geo mean counts, and hepatopulmonary shunt fraction

Figure 2 Software algorithm shows the entire process of hepatopulmonary shunt calculation by this software

Software code

This software has various sections of code, i.e., variable declaration, image validation, main processing, and HPS calculation.

Variable declaration

In this section of code, all the variables are declared to store image, region of interest (ROI) counts and pixels.

Image validation

This part of code validates the images in runtime during the launch of the software. It only permits defined image type and format, which is validated by this section of code; otherwise, validation fails and error massage is displayed. Two sets of images in anterior and posterior projections, i.e., total four static images are validated. These images are then transferred to the main processing section.

Main processing

In this part of code, images are loaded in the software and displayed in the first two rows. Once all four images are displayed, software prompts to draw ROI on the first set of images, and the same ROIs are copied on the second set of images and ROI analysis takes place. Total ROI counts and total ROI pixels are captured and stored in the declared variable. Pixel normalized background corrected counts are obtained for right lung, left lung and liver ROIs using Equation 1 for anterior as well as posterior projections. Geometric mean count is also calculated for right lung, left lung, and liver using Equation 2.

Hepatopulmonary shunt calculation

This part of code is designed to calculate HPS using Equation 3.

Software installation

This software named HPS can be installed on all versions of Xeleris workstations using Aladdin editor with slight modifications.

Where:

Ccor= Background corrected count

Croi= Total ROI count

Cbkg= Total background count

Proi= Total ROI pixel

Pbkg= Total background pixel

Where:

Cgeo= geometric mean count

Cant= Anterior count

Cpost= Posterior count

Where:

% HPS = Hepatopulmonary shunt

Cliv= Total liver mean count (background corrected)

ClLung= Total left lung mean count (background corrected)

CrLung= Total right Lung mean count (background corrected)

Imaging

Static and SPECT imaging was performed on Infinia Hawkeye 1.123, GE medical system, Haifa, Israel, after transarterial administration of freshly prepared 370MBq99m Tc-MAA in the liver tumor. Two sets of static images in anterior and posterior projection and one set SPECT were performed including lung and liver in the same frame. Those two sets of static images-one with a Co-57 flood phantom SET-1 and another without the phantom SET-2 are acquired for 10 min each on gamma camera in anterior and posterior projection in 256 × 256 matrix and 1 zoom including lung and liver in the same field of view in identical position.

Imaging requirement for clinical use of software

Two sets of static images are required for processing on this software. Name of the images should be preset as Ant-HPS-Ph and Post-HPS-Ph for SET-1 anterior and posterior images and Ant-HPS and Post-HPS for SET-2 anterior and posterior images.

Image processing and hepatopulmonary shunt calculation by this software

All four images from two image sets are selected and opened in HPS Software on XELERIS workstation. The software opens with active main button and no image loaded in it. Once the main button is clicked both the sets of images are validated and loaded in the image cell of the first two row of image container [Figure 1]. First row contains SET-1 image and second row contains SET-2 images and third row remains empty. Software prompts to draw right lung, left lung, liver, and three respective background ROIs on both projections of SET-1 images. Once all the ROIs are drawn HPS button becomes active. HPS can be calculated by clicking HPS button on the left side panel. The background corrected left and right lung and liver counts are displayed in the 5th cell of image container and HPS fraction is displayed in the 6th cell of image container [Figure 1].

Software validation

Debugging

This software was evaluated for its robustness and accuracy in estimating HPS before routine clinical utilization of this software. Debugging of software is performed for all the involved processes such as image validation, ROI generation, and HPS calculation.

Phantom-based validation

Phantom preparation

Two phantoms, i.e., NEMA IQ Phantom (Data Spectrum Inc., Durham, NC, USA) and Jaszczak phantom (Data Spectrum Inc., Durham, NC, USA) were used to make HPS Reference Model (HRM) phantom [Figure 3]. IQ phantom was used to mimic lung, and Jaszczak phantom was used to mimic liver in HRM phantom. HRM Phantom was filled with99m TcO4− to create model for HPS 5%, 10%, 15%, 20%, and 30% by filling up IQ and Jaszczak phantom in respective ratio as described in [Table 1].

Figure 3 Hepatopulmonary shunt reference model phantom

Table 1 Activity in the National Electrical Manufacturers Association image quality and Jaszczak single-photon emission computed tomography phantom mimicking lung and liver, respectively

Phantom imaging

Phantom imaging was performed in the clinical parameters as described above twice in respective HPS settings in test and retest setting for 2 min per acquisition.

Clinical validation

HPS imaging data of all the patients referred to our department for HPS study were utilized for validation of this software. The study was approved by intuitional ethics committee. Software validation was performed by comparing the HPS calculated by the automated software with HPS calculated by manual method.

Statistical test

Phantom study

The percentage error in HPS estimation by software was calculated for test-retest setting. Furthermore, the percentage error in HPS estimation by software with respect to the reference HPS value was calculated.

Clinical study

Average and standard deviation were calculated for HPS obtained by software and manual method as well. The correlation coefficient was calculated between HPS obtained by Software and manual method. Wilcoxon Signed-Rank Test was also performed to find the significance between the HPS calculated by this software and manual method.

Results

This software passed all the debugging tests. For example, if valid image sets are not available user is prompted to select valid images. It prompts the user to draw specific ROIs on a specific image and does not allow drawing ROI on wrong image set. Automatically, ROIs are copied on SET-2 images utilized for ROI analysis. It is able to fetch ROI data such as total counts in ROI, average counts in ROI, and number of pixels in ROI for further calculation.

The HPS estimation performed in test-retest and compared [Figure 4]. The percentage error in HPS estimated by this software in test-retest setting across all the HPS level was 2.07% [Table 2]. Furthermore, the percentage error in HPS estimation in comparison with that of reference HPS across all the HPS level was 5.20% [Table 2].

Figure 4 Hepatopulmonary shunt calculated by hepatopulmonary shunt software using in-house prepared hepatopulmonary shunt phantom

Table 2 The percentage error in hepatopulmonary shunt estimation by hepatopulmonary shunt software in comparison reference percentage with at various hepatopulmonary shunt leve

A total of 36 patient's (35 males and 1 female), data were processed and HPS was calculated using this software and manual method. There was no much difference noted in HPS calculated by software HPS calculated by manual method [Table 3]. HPS of both the sets were compared, we found excellent correlation between two groups (correlation coefficient = 0.998). Wilcoxon Signed-Rank Test was performed, no significant difference at P ≤ 0.01 was found between two sets of HPS calculated by Software and manual method (P = 0.15272). The [Table 4] shows the details result of Wilcoxon Signed-Rank Test.

Table 3 Hepatopulmonary shunt calculated by software and manual method

Table 4 Wilcoxon Signed-Rank test result

Discussion

HPS in TARE therapy is widely reported in the literature resulting in fractional accumulation of microsphere in the lung and an unwanted radiation dose to the lung. As per literature the radiation dose >30 Gy as single exposure and >50 Gy cumulative exposure to the lungs may cause irreversible damage to the lung.

Hence, radiation exposure to the lung due to HPS is one of the limiting factors to perform TARE.[12],[13],[14],[15]

As per available literature, reduction of administered therapeutic activity of 90Y-microsphere is suggested for the patients with HPS between 20% and 30% whereas no dose modification is suggested below 20% of HPS. However, for patients with above 30% HPS, TARE is not recommended, considering radiotoxicity to the lungs.[6],[7],[10],[11],[12],[13],[14],[15] Hence, accurate estimation of HPS is mandatory to avoid any radiation-related complication to the patient. HPS is estimated in the Nuclear Medicine Department by gamma camera imaging after transarterial administration of99m Tc-MAA in the interventional radiology department. Most of the vendors supplying gamma camera or SPECT systems are not providing software for HP Shunt calculation, hence manual calculation of HPS is the only option left with the nuclear medicine professionals. With the manual technique, drawing liver and lung ROI accurately become difficult without anatomical demarcations which may cause erroneous HPS calculations with harmful implications. A precisely drawn ROI and proper background correction technique are mandatory to calculate accurate HPS. In our HPS Software, the Co-57 flood phantom image is used for accurate organ delineation and also helps to draw precise ROI around lung, liver, and background [Figure 1].

Our software uses an established technique of pixel normalized background correction, and geometric mean techniques to obtain accurate counts from the ROIs and HP shunt formula (Equation 3) is used for HPS calculation. HPS Software prompts the operator on every stage of image processing which makes this software user-friendly and reliable for HPS estimation. This software was also tested for the robustness. Software shows excellent result with the phantom as well as clinical studies. Error in HPS estimation in test-retest was found to be as low as 2%, and for test-reference, it was around 5%. We observed good correlation between the HPS calculated by manual method and software method in clinical study. Our results with phantom study as well as clinical study show good reproducibility, reliability, and robustness of this software.

Conclusion

Liver-lung interface is better delineated with the incorporation of phantom image; enabling accurate ROI placement which results in robust calculation of HPS. The clinical impact is seen in large tumors which obscure the liver margins. Self-instructional nature of the software makes it user-friendly.

Clinical relevance

Accurate estimation of HPS should be done as it is a dose-limiting factor in TARE. HPS software is a precise tool to estimate HPS accurately.

Conflict of Interest

There are no conflicts of interest.

Financial support and sponsorship

Nil.

• References

• 1 Owunwanne A, Patel M, Sadek S. Preparation, technetium-99m macroaggregated albumin. Handbook of Radiopharmaceuticals. Chapman and Hall; 1995. p. 67.
• 2 Murthy R, Nunez R, Szklaruk J, Erwin W, Madoff DC, Gupta S, et al. Yttrium-90 microsphere therapy for hepatic malignancy: Devices, indications, technical considerations, and potential complications. Radiographics 2005;25 Suppl 1:S41-55.
• 3 Salem R, Lewandowski RJ, Atassi B, Gordon SC, Gates VL, Barakat O, et al. Treatment of unresectable hepatocellular carcinoma with use of 90Y microspheres (TheraSphere): Safety, tumor response, and survival. J Vasc Interv Radiol 2005;16:1627-39.
• 4 Salem R, Thurston KG. Radioembolization with yttrium-90 microspheres: A state-of-the-art brachytherapy treatment for primary and secondary liver malignancies: Part 3: Comprehensive literature review and future direction. J Vasc Interv Radiol 2006;17:1571-93.
• 5 Kennedy A, Nag S, Salem R, Murthy R, McEwan AJ, Nutting C, et al. Recommendations for radioembolization of hepatic malignancies using yttrium-90 microsphere brachytherapy: A consensus panel report from the radioembolization brachytherapy oncology consortium. Int J Radiat Oncol Biol Phys 2007;68:13-23.
• 6 TheraSphere™ yttrium-90 Glass Microspheres Instructions for use. Surrey, UK: Biocompatibles UK Ltd.; 2005. Available from: http://www.therasphere.com/physicians-packageinsert/TS_PackageInsert_USA_v12.pdf. [Last accessed on 2017 Jan 25].
• 7 Sirtex Medical Limited. Sirtex Medical Training Manual, Training Program Physicians and Institutions. Australia: Sirtex Medical Limited; 2015. Available from: http://foxfireglobal.sirtex.com/sites/foxfireglobal.sirtex.com/files/user/trn-rw-04_for_eu_au_nz_and_asia.pdf. [Last accessed on 2017 Jan 25].
• 8 Ziessman HA, Thrall JH, Yang PJ, Walker SC, Cozzi EA, Niederhuber JE, et al. Hepatic arterial perfusion scintigraphy with Tc-99m-MAA. Use of a totally implanted drug delivery system. Radiology 1984;152:167-72.
• 9 Jha AK, Zade AA, Rangarajan V, Purandare N, Shah SA, Agrawal A, et al. Comparative analysis of hepatopulmonary shunt obtained from pretherapy 99mTc MAA scintigraphy and post-therapy 90Y bremsstrahlung imaging in 90Y microsphere therapy. Nucl Med Commun 2012;33:486-90.
• 10 Jha AK, Purandare N, Shah SA, Agrawal A, Rangarajan V, Kulkarni SS. Is there a correlation between planar scintigraphy after 99mTc-MAA and 90Y administration? Nucl Med Commun 2016;37:107-9.
• 11 Leung TW, Lau WY, Ho SK, Ward SC, Chow JH, Chan MS, et al. Radiation pneumonitis after selective internal radiation treatment with intraarterial 90yttrium-microspheres for inoperable hepatic tumors. Int J Radiat Oncol Biol Phys 1995;33:919-24.
• 12 Yip D, Allen R, Ashton C, Jain S. Radiation-induced ulceration of the stomach secondary to hepatic embolization with radioactive yttrium microspheres in the treatment of metastatic colon cancer. J Gastroenterol Hepatol 2004;19:347-9.
• 13 Goin JE, Salem R, Carr BI, Dancey JE, Soulen MC, Geschwind JF, et al. Treatment of unresectable hepatocellular carcinoma with intrahepatic yttrium 90 microspheres: Factors associated with liver toxicities. J Vasc Interv Radiol 2005;16:205-13.
• 14 Szyszko T, Al-Nahhas A, Tait P, Rubello D, Canelo R, Habib N, et al. Management and prevention of adverse effects related to treatment of liver tumours with 90Y microspheres. Nucl Med Commun 2007;28:21-4.
• 15 Denecke T, Rühl R, Hildebrandt B, Stelter L, Grieser C, Stiepani H, et al. Planning transarterial radioembolization of colorectal liver metastases with yttrium 90 microspheres: Evaluation of a sequential diagnostic approach using radiologic and nuclear medicine imaging techniques. Eur Radiol 2008;18:892-902.

Publikationsverlauf

Eingereicht: 17. Dezember 2018

Angenommen: 18. Mai 2019

Artikel online veröffentlicht:
22. April 2022