Background

Selective internal radiation therapy (SIRT) with direct 99Y-microspheres infusion into the hepatic artery is a novel, promising method for palliative treatment of primary and metastatic liver cancer [1, 2]. The most important problem associated with this therapy is to calculate an adequate activity of 99Y-microspheres, which can destroy liver tumors while sparing healthy liver tissue [3, 4]. The 99mTc-MAA scans are performed in order to calculate a 99Y dose and to predict a post infusion 99Y-microspheres distribution profile between the liver parenchyma and tumor compartments (the T/N ratio). The post procedural 99Y dose estimation in liver is very difficult because direct measurement of β particles is not possible with SPECT/CT. Thus, 99Y bremsstrahlung single-photon emission computed tomography/computed tomography (SPECT/CT) is performed to assess post-treatment 99Y-microspheres distribution [3, 5]. In reality, there are two ways to calculate dosimetry after SIRT. The first one is the MIRD equation based on the assumption of uniform 99Y distribution between the tumor and liver tissue compartments [6]. The second one is Monte Carlo simulation in which SPECT/CT images are used to obtain the isodoses curves and histograms of the target liver lesions [7]. For improvement of SIRT treatment results a detailed knowledge about 99Y-microspheres distribution within the liver is required [3]. This way patients prone to potential radiation-induced liver disease (RILD) or patients with non-curative tumor dose may be selected immediately after a SIRT procedure. In the present paper, we assessed the 90Y-PET/CT method as an alternative for the 99nlTc-MAA and 90Y-SPECT (bremsstrahlung) images in clinical studies.

Material and methods

Patient studies

Three patients (2 males, 1 female) were treated with 90Y microspheres selective internal radioembolisation. All patients with unresectable liver metastases had adequate performance status and acceptable liver and renal function. Two patients had colorectal liver metastasis (mCRC) and one had metastasis from pancreatic cancer. A total of nine tumors was evaluated, 7 mCRC and 2 pancreatic cancer cases. The study was approved by Ethics Committee. All the patients were included in the study after signing a written informed consent.

Image evaluation

The patients were assessed before and after selective internal radioembolisation (SIRT) with computed tomography (CT). The whole liver volume, the liver tumors longest diameter and volume were collected. The 90Y-PET studies were performed between 8 and 48 hours after resin microspheres administration into the liver. The Response Evaluation Criteria in Solid Tumors 1.1 (RECIST) were used to evaluate liver tumors response [8]. The correlations between the target liver lesions detected with 90Y-PET, 90Y-SPECT and 99mTc-MAA were investigated and compared with CT images obtained one month before and two or three months after procedure. The tumor to normal liver tissue ratio (T/N ratio) in 99mTc-MAA, in 90Y-SPECT and in 90Y-PET for each selected lesion was calculated. Based on the data, the absorbed dose for the liver parenchyma and for each target liver tumor was calculated using the internal dosimetry schema of the Medical internal Dose (MIRD) Committee of Society of Nuclear Medicine [6]. The absorbed dose was calculated using the T/N ratio in 99Tc-MAA, in 90Y-bremsstrahlung and in 90Y-PET. The target lesion in the liver was described as a lesion with the longest diameter of minimum 10 mm and clearly visible on CT, 90Y-PET, 90Tc MAA SPECT and 90Y-SPECT. A GE Discovery ST PET/CT scanner with a copper ring protected the gantry was used for images acquisition. The ring thickness was 2 mm. The ring width was sufficient to cover the area of the detector (and equal to the axial FOV). The time of a single 90Y-PET scan was 20 min (one patient) and 30 min (two patients). The role of the copper ring was to absorb the bremsstrahlung photons and to prevent saturations of the detectors. A GE Infinia VCHWK4 with HEPG collimators was used for SPECT/CT imaging. The energy window was 140 keV ± 100%.

SIRT procedure

The radioembolization procedure was conducted in line with guidelines approved by panel experts [3, 5].

Candidates for radioembolization therapy were qualified by a multidisciplinary team consisting of interventional radiologists, oncologists, nuclear medicine specialists and surgeons in line with strict inclusion and exclusion criteria, which were published previously [3, 5].

During the pre-treatment procedure, selective coll embolization of the gastroduodenal artery, right gastric artery and gallbladder artery were performed. The hepatopulmonary shunt, a potential gastrointestinal leak and the tumor to liver ratio were assessed after SPECT 99mTc-MAA. An activity of 4mci 99Tc MAA was administered into the hepatic artery. The Body Surface Area Method was used for the patients’ 99Y-microspheres dose calculation [5]. All the patients were treated with 99Y biocompatible, not biodegradable SIR-Spheres microspheres (Sirtex Medical inc., Australia). The 99Y is a pure beta emitter with a liver tissue penetration of 2.5 mm and average energy of 0.94 MeV and a half-life of 2.67 days. The average resin microspheres’ diameter is 35 ± 5 mm [3, 5]. The SIRT procedure was performed after superselectlve catheterization of hepatic artery branches by a slow, controlled 99Y microspheres injection. All the patients received whole liver treatment. After the therapy, bremsstrahlung with SPECT and 99Y-PET images were made to check 99Y-microspheres liver dose deposition.

Results

All SIRT procedures were made with technical success. The whole dose prescribed for each patient was administered. There were no serious adverse events (SAE) observed. After the therapy, two patients had transient nausea, vomiting and mild pain in the liver region which required only symptomatic treatment. The details of clinical and treatment data are summarized in Table 1.

A total of nine tumor target lesions were evaluated with imaging methods. Taking into consideration the response rate for each solid tumor, two CR, two PR and three SD were observed for seven mCRC at first follow-up. For two pancreatic lesions SD was found. The 90Y absorbed dose in the liver tissue ranged from 8.5 Gy to 25.8 Gy (mean 15.5 Gy). The mean T/N ratio was 2.7 in 99mTc-MAA, 2.3 in 90Y-SPECT and 3.6 in 90Y-PET. The mean 90Y absorbed dose in tumor was 133 Gy, 112 Gy, and 187 Gy, respectively. All the tumors except one had sufficient (> 70 Gy) absorbed dose calculated using the T/N ratio revealed in imaging. Only one mCRC lesion (No 2) had absorbed dose lower than 70 Gy (based on MAA T/N ratio). In this case a complete response was confirmed in the first follow-up. The image response parameters, the T/N ratio and the estimated (MiRD) dose calculations for each tumor are summarized in Table 2.

Discussion

In the recent years, there has been agrowing interest observed in the use of 90Y microspheres for regional liver tumor therapy. Promising treatment results such as improvements of time to progression, median overall survival and safety profile are an encouragement for further development of this method [1, 9-11]. Although we are aware that one of the m ain i im itations of our study is a small group of patients, we can also confirm high efficiency of this therapy. We noticed one partial response and two stable diseases in our study. The time to progression ranged from 117 to 217 days in colorectal metastases and 214 days in pancreatic cancer metastases. There are no serious adverse events associated with the therapy. In our opinion, there is still room for improvement of results in this treatment. The first way is to treat patients as soon as possible. At present, results of prospective randomized multicenter trials are expected (i.e. SIRFLOX), in which radioembolisation in first line treatment is evaluated [12]. The second way is to improve knowledge on 90Y-microspheres liver distribution, especially 90Y liver dosimetry. The background for intra-arterial 90Y-microspheres therapy is a special type of vascular anatomy of the liver. The majority of liver tumors’ blood supply originates from the hepatic artery branches, then the portal vein (about 80-100% of their supply for tumors >3 mm) [13]. It is estimated that a dose higher than 70 Gy is required to destroy most of the liver tumors. High 90Y-microspheres concentration within liver cancer leads to their destruction by way of ionizing radiation and embolization. The former has a crucial importance for the therapy [14]. The main limitation of radioembolisation is low liver tissue tolerance to radiation, with a possibility of serious adverse events after irradiation higher than 30 Gy [14, 15]. In our study, the radiation of the liver was estimated from 8.4 Gy to 25.8 Gy (mean 15.5 Gy) based on the MIRD formula and no serious side effects were reported (i.e. RILD). To predict post-infusion 90Y-microspheres’ distribution within the liver parenchyma 99mTc-MAA scans are taken. Nevertheless, the real distribution of the 90Y-microspheres might be different from the 99mTc-MAA because of a higher albumin diameter and intra-hepatic blood flow variations [3, 5, 16]. After the therapy, 90Y-bremsstrahlung single-photon emission computed tomography/computed tomography (SPECT/CT) is performed to evaluate the real distribution of 90Y-microspheres [3, 5]. Unfortunately insufficient spatial resolution of 90Y-bremsstrahlung images is a main limitation of this method and can lead to inaccurate evaluation of 90Y-microspheres distribution [17]. To evaluate intra-hepatic 90Y administration some authors use a 99mTc-MAA injection immediately after a SIRT procedure [18]. However, in our opinion, embolic effects of resin microspheres may lead to inaccurate 99mTc-MAA deposition in the liver and conclusions based on this method may therefore be misleading. A novel approach to assessment of radioembolisation effects consists in taking 90Y PET images [19, 20]. Although 90Y is traditionally considered a pure β-emitter, its decay has a minor branch to the 0+ first excited state of 90Zr at 1.78 MeV. De-excitation consists in emission of either a conversion electron or an internal e^-e+ pair creation. It happens in 32 out of 1 million decays and might be imaged with PET [20, 21]. We assessed the 90Y-PET method as an alternative for the99mTc-MAA and bremsstrahlung SPECT/CT images in clinical studies. We performed a 90Y-PET test between 8 to 24 hours after SiRT and we did not observe deterioration of image quality over time. As the 90Y half-life is 2.67 days, the PET scan should be taken within this time limit in our opinion [3, 13, 21]. The time of a single PET imaging was 20 min in one case and 30 min in the next two patients. To obtain better image results we suggest taking a PET scan lasting 30 min for an administered dose ranging 1.2-2.2 GBq of 90Y-microspheres. In order to protect the gantry detectors, we used a copper ring to absorb photons of low energy and we observed that 2 mm thickness of the ring is sufficient to achieve this aim. However, some authors do not use any protection and they have not noticed its saturation with total activity of 2.0 GBq. Yet, sometimes a higher dose may be needed for a patient’s treatment and we suggest more caution in such cases [22].

We noticed a similar T/N ratio calculated using 99mTc-MAA and bremsstrahlung SPECT/CT images for most liver tumors and considerable differences when we compare them with the 90Y-PET-revealed T/N ratio. The mean T/N ratio was 2.7 in 99mTc-MAA, 2.3 in 90Y-SPECT and 3.6 in 90Y-PET. The mean 90Y absorbed dose in tumor was 133 Gy, 112 Gy, and 187 Gy respectively. It leads also to significant differences between the calculated 90Y liver tumors’ absorbed doses. We can see doses, which are 2-3 times higher when estimated with the 90Y-PET T/N ratio. It means that the dose required to destroy the tumor (approximately 70-90 Gy) is significantly exceeded [14, 15]. According to our study tumor absorbed doses calculated with 90Y-PET ranged from 144 Gy to 283 Gy, but when we consider tumor treatment results we can see mostly stable disease in overall response rate. Therefore, the absorbed doses calculated on the basis of the 99mTc-MAA and bremsstrahlung SPECT/CT ratio ranging from 64 Gy to 183 Gy (except tumor No. 8) seem to be more close to reality. One of the explanations of these results may be lower sensitivity of 90Y-PET/CT compared to 90Y-bremsstrahlung [20]. Thus, to assess the real T/N ratio based on 90Y-PET/CT, further studies are necessary (Figure 1). On the other hand, its resolution and contrast are much better and 90Y-PET/CT imaging may be used not only to detect liver lesions after radioembolisation, but also to distinguish a necrosis area within solid tumors. Better resolution of this method may help to find out leaks of 90Y microspheres to the digestive tract in our opinion [17, 20]. From our point of view, the possibility of estimation of radiation dose in liver lesions is very interesting.

Figure 1. The same liver lesion in CT, 99mTc-MAA, 90Y-bremsstrahlung and 90Y-PET scans

Conclusion

90Y-PET/CT images give crucial information regarding 90Y-microspheres distribution and dosimetry and may serve as a predictor of efficiency of radioembolisation.

Correspondence to:
Piotr Piasecki, MD
Interventional Radiology Department
of Military Institute of Medicine
ul. Szaserów 128, 04-141 Warsaw, Poland
E-mail: piotr_piasecki@wp.pl