SPECT image segmentation for estimation of tumour volume and activity concentration in 177Lu-DOTATATE radionuclide therapy

A review of 177Lu dosimetry workflows: how to reduce the imaging workloads?

177 Lu radiopharmaceutical therapy is a standardized systemic treatment, with a typical dose of 7.4 GBq per injection, but its response varies from patient to patient. Dosimetry provides the opportunity to personalize treatment, but it requires multiple post-injection images to monitor the radiopharmaceutical's biodistribution over time. This imposes an additional imaging burden on centers with limited resources. This review explores methods to lessen this burden by optimizing acquisition types and minimizing the number and duration of imaging sessions. After summarizing the different steps of dosimetry and providing examples of dosimetric workflows for177 Lu -DOTATATE and177 Lu -PSMA, we examine dosimetric workflows based on a reduced number of acquisitions, or even just one. We provide a non-exhaustive description of simplified methods and their assumptions, as well as their limitations. Next, we detail the specificities of each normal tissue and tumors, before reviewing dose-response relationships in the literature. In conclusion, we will discuss the current limitations of dosimetric workflows and propose avenues for improvement.

Pareto optimization of SPECT acquisition and reconstruction settings for 177Lu activity quantification

BACKGROUND

The aim was to investigate the noise and bias properties of quantitative 177Lu-SPECT with respect to the number of projection angles, and the number of subsets and iterations in the OS-EM reconstruction, for different total acquisition times.

METHODS

Experimental SPECT acquisition of six spheres in a NEMA body phantom filled with 177Lu was performed, using medium-energy collimators and 120 projections with 180 s per projection. Bootstrapping was applied to generate data sets representing acquisitions with 20 to 120 projections for 10 min, 20 min, and 40 min, with 32 noise realizations per setting. Monte Carlo simulations were performed of 177Lu-DOTA-TATE in an anthropomorphic computer phantom with three tumours (2.8 mL to 40.0 mL). Projections representing 24 h and 168 h post administration were simulated, each with 32 noise realizations. Images were reconstructed using OS-EM with compensation for attenuation, scatter, and distance-dependent resolution. The number of subsets and iterations were varied within a constrained range of the product number of iterations × number of projections ≤ 2400 . Volumes-of-interest were defined following the physical size of the spheres and tumours, the mean activity-concentrations estimated, and the absolute mean relative error and coefficient of variation (CV) over noise realizations calculated. Pareto fronts were established by analysis of CV versus mean relative error.

RESULTS

Points at the Pareto fronts with low CV and high mean error resulted from using a low number of subsets, whilst points at the Pareto fronts associated with high CV but low mean error resulted from reconstructions with a high number of subsets. The number of projection angles had limited impact.

CONCLUSIONS

For accurate estimation of the 177Lu activity-concentration from SPECT images, the number of projection angles has limited importance, whilst the total acquisition time and the number of subsets and iterations are parameters of importance.

Lutetium-177-Prostate-Specific Membrane Antigen Radioligand Therapy: What Is the Value of Post-Therapeutic Imaging?

Lutetium-177 (Lu-177)-labelled radioligand therapies (RLT) targeting prostate-specific membrane antigen (PSMA) present a promising treatment for patients with progressive metastasized castration-resistant prostate cancer (mCRPC). Personalized dosimetry, facilitated by post-therapeutic imaging, offers the potential to enhance treatment efficacy by customizing radiation doses to individual patient needs, thereby maximizing therapeutic benefits while minimizing toxicity to healthy tissues. However, implementing personalized dosimetry is resource-intensive, requiring multiple single-photon emission-computed tomography (SPECT)/CT scans and posing significant logistical challenges for both healthcare facilities and patients. Despite these challenges, personalized dosimetry can lead to optimized radiation delivery, improved safety, and better management of complex cases. Nevertheless, the financial and resource burdens complicate its adoption in routine clinical practice. While the European Association of Nuclear Medicine (EANM) supports personalized dosimetry, standardization is lacking due to these practical constraints. Further research and streamlined methodologies are essential to balance the benefits and feasibility of personalized dosimetry, potentially improving treatment outcomes for mCRPC patients.

Relationship Between Absorbed Dose and Response in Neuroendocrine Tumors Treated with [177Lu]Lu-DOTATATE

Peptide receptor radionuclide therapy presents the possibility of tracing and quantifying the uptake of the drug in the body and performing dosimetry, potentially allowing individualization of treatment schemes. However, the details of how neuroendocrine tumors (NETs) respond to different absorbed doses are insufficiently known. Here, we investigated the relationship between tumor-absorbed dose and tumor response in a cohort of patients with NETs treated with [177Lu]Lu-DOTATATE. Methods: This was a retrospective study based on 69 tumors in 32 patients treated within a clinical trial. Dosimetry was performed at each cycle of [177Lu]Lu-DOTATATE, rendering 366 individual absorbed dose assessments. Hybrid planar-SPECT/CT imaging using [177Lu]Lu-DOTATATE was used, including quantitative SPECT reconstruction, voxel-based absorbed dose rate calculation, semiautomatic image segmentation, and partial-volume correction. Changes in tumor volume were used to determine tumor response. The volume for each tumor was manually delineated on consecutive CT scans, giving a total of 712 individual tumor volume assessments. Tumors were stratified according to grade. The relationship between absorbed dose and response was investigated using mixed-effects models and logistic regression. Tumors smaller than 4 cm3 were excluded. Results: In grade 2 NETs, a clear relationship between absorbed dose and volume reduction was observed. Our observations suggest a 90% probability of partial tumor response for an accumulated tumor-absorbed dose of at least 135 Gy. Conclusion: Our findings are in accordance with previous observations regarding the relationship between tumor shrinkage and absorbed dose. Moreover, our data suggest an absorbed dose threshold for partial response in grade 2 NETs. These observations provide valuable insights for the design of dosimetry-guided peptide receptor radionuclide therapy schemes.

3D printed non-uniform anthropomorphic phantoms for quantitative SPECT

BACKGROUND

A 3D printing grid-based method was developed to construct anthropomorphic phantoms with non-uniform activity distributions, to be used for evaluation of quantitative SPECT images. The aims were to characterize the grid-based method and to evaluate its capability to provide realistically shaped phantoms with non-uniform activity distributions.

METHODS

Characterization of the grid structures was performed by printing grid-filled spheres. Evaluation was performed by micro-CT imaging to investigate the printing accuracy and by studying the modulation contrast ([Formula: see text]) in SPECT images for 177Lu and 99mTc as a function of the grid fillable-volume fraction (FVF) determined from weighing. The grid-based technique was applied for the construction of two kidney phantoms and two thyroid phantoms, designed using templates from the XCAT digital phantoms. The kidneys were constructed with a hollow outer container shaped as cortex, an inner grid-based structure representing medulla and a solid section representing pelvis. The thyroids consisted of two lobes printed as grid-based structures, with void hot spots within the lobes. The phantoms were filled with solutions of 177Lu (kidneys) or 99mTc (thyroids) and imaged with SPECT. For verification, Monte Carlo simulations of SPECT imaging were performed for activity distributions corresponding to those of the printed phantoms. Measured and simulated SPECT images were compared qualitatively and quantitatively.

RESULTS

Micro-CT images showed that printing inaccuracies were mainly uniform across the grid. The relationships between the FVF from weighing and [Formula: see text] were found to be linear (r = 0.9995 and r = 0.9993 for 177Lu and 99mTc, respectively). The FVF-deviations from the design were up to 15% for thyroids and 4% for kidneys, mainly related to possibilities of cleaning after printing. Measured and simulated SPECT images of kidneys and thyroids exhibited similar activity distributions and quantitative comparisons agreed well, thus verifying the grid-based method.

CONCLUSIONS

We find the grid-based technique useful for the provision of 3D printed, realistically shaped, phantoms with non-uniform activity distributions, which can be used for evaluation of different quantitative methods in SPECT imaging.

The Challenge of Single-Photon Emission Computed Tomography Image Segmentation in the Internal Dosimetry of 177Lu Molecular Therapies

The aim of this article is to review the single photon emission computed tomography (SPECT) segmentation methods used in patient-specific dosimetry of 177Lu molecular therapy. Notably, 177Lu-labelled radiopharmaceuticals are currently used in molecular therapy of metastatic neuroendocrine tumours (ligands for somatostatin receptors) and metastatic prostate adenocarcinomas (PSMA ligands). The proper segmentation of the organs at risk and tumours in targeted radionuclide therapy is an important part of the optimisation process of internal patient dosimetry in this kind of therapy. Because this is the first step in dosimetry assessments, on which further dose calculations are based, it is important to know the level of uncertainty that is associated with this part of the analysis. However, the robust quantification of SPECT images, which would ensure accurate dosimetry assessments, is very hard to achieve due to the intrinsic features of this device. In this article, papers on this topic were collected and reviewed to weigh up the advantages and disadvantages of the segmentation methods used in clinical practice. Degrading factors of SPECT images were also studied to assess their impact on the quantification of 177Lu therapy images. Our review of the recent literature gives an insight into this important topic. However, based on the PubMed and IEEE databases, only a few papers investigating segmentation methods in 177Lumolecular therapy were found. Although segmentation is an important step in internal dose calculations, this subject has been relatively lightly investigated for SPECT systems. This is mostly due to the inner features of SPECT. What is more, even when studies are conducted, they usually utilise the diagnostic radionuclide 99mTc and not a therapeutic one like 177Lu, which could be of concern regarding SPECT camera performance and its overall outcome on dosimetry.

Feasibility of 177Lu activity quantification using a small portable CZT-based gamma-camera

BACKGROUND

In image processing for activity quantification, the end goal is to produce a metric that is independent of the measurement geometry. Photon attenuation needs to be accounted for and can be accomplished utilizing spectral information, avoiding the need of additional image acquisitions. The aim of this work is to investigate the feasibility of 177Lu activity quantification with a small CZT-based hand-held gamma-camera, using such an attenuation correction method.

METHODS

A previously presented dual photopeak method, based on the differential attenuation for two photon energies, is adapted for the three photopeaks at 55 keV, 113 keV, and 208 keV for 177Lu. The measurement model describes the count rates in each energy window as a function of source depth and activity, accounting for distance-dependent system sensitivity, attenuation, and build-up. Parameter values are estimated from characterizing measurements, and the source depth and activity are obtained by minimizing the difference between measured and modelled count rates. The method is applied and evaluated in phantom measurements, in a clinical setting for superficial lesions in two patients, and in a pre-clinical setting for one human tumour xenograft. Evaluation is made for a LEHR and an MEGP collimator.

RESULTS

For phantom measurements at clinically relevant depths, the average (and standard deviation) in activity errors are 17% ± 9.6% (LEHR) and 2.9% ± 3.6% (MEGP). For patient measurements, deviations from activity estimates from planar images from a full-sized gamma-camera are 0% ± 21% (LEHR) and 16% ± 18% (MEGP). For mouse measurements, average deviations of - 16% (LEHR) and - 6% (MEGP) are obtained when compared to a small-animal SPECT/CT system. The MEGP collimator appears to be better suited for activity quantification, yielding a smaller variability in activity estimates, whereas the LEHR results are more severely affected by septal penetration.

CONCLUSIONS

Activity quantification for 177Lu using the hand-held camera is found to be feasible. The readily available nature of the hand-held camera may enable more frequent activity quantification in e.g., superficial structures in patients or in the pre-clinical setting.

Semi-automated segmentation methods of SSTR PET for dosimetry prediction in refractory meningioma patients treated by SSTR-targeted peptide receptor radionuclide therapy

OBJECTIVES

Tumor dosimetry with somatostatin receptor-targeted peptide receptor radionuclide therapy (SSTR-targeted PRRT) by 177Lu-DOTATATE may contribute to improved treatment monitoring of refractory meningioma. Accurate dosimetry requires reliable and reproducible pretherapeutic PET tumor segmentation which is not currently available. This study aims to propose semi-automated segmentation methods to determine metabolic tumor volume with pretherapeutic 68Ga-DOTATOC PET and evaluate SUVmean-derived values as predictive factors for tumor-absorbed dose.

METHODS

Thirty-nine meningioma lesions from twenty patients were analyzed. The ground truth PET and SPECT volumes (VolGT-PET and VolGT-SPECT) were computed from manual segmentations by five experienced nuclear physicians. SUV-related indexes were extracted from VolGT-PET and the semi-automated PET volumes providing the best Dice index with VolGT-PET (Volopt) across several methods: SUV absolute-value (2.3)-threshold, adaptative methods (Jentzen, Otsu, Contrast-based method), advanced gradient-based technique, and multiple relative thresholds (% of tumor SUVmax, hypophysis SUVmean, and meninges SUVpeak) with optimal threshold optimized. Tumor-absorbed doses were obtained from the VolGT-SPECT, corrected for partial volume effect, performed on a 360° whole-body CZT-camera at 24, 96, and 168 h after administration of 177Lu-DOTATATE.

RESULTS

Volopt was obtained from 1.7-fold meninges SUVpeak (Dice index 0.85 ± 0.07). SUVmean and total lesion uptake (SUVmeanxlesion volume) showed better correlations with tumor-absorbed doses than SUVmax when determined with the VolGT (respective Pearson correlation coefficients of 0.78, 0.67, and 0.56) or Volopt (0.64, 0.66, and 0.56).

CONCLUSION

Accurate definition of pretherapeutic PET volumes is justified since SUVmean-derived values provide the best tumor-absorbed dose predictions in refractory meningioma patients treated by 177Lu-DOTATATE. This study provides a semi-automated segmentation method of pretherapeutic 68Ga-DOTATOC PET volumes to achieve good reproducibility between physicians.

CLINICAL RELEVANCE STATEMENT

SUVmean-derived values from pretherapeutic 68Ga-DOTATOC PET are predictive of tumor-absorbed doses in refractory meningiomas treated by 177Lu-DOTATATE, justifying to accurately define pretherapeutic PET volumes. This study provides a semi-automated segmentation of 68Ga-DOTATOC PET images easily applicable in routine.

KEY POINTS

• SUVmean-derived values from pretherapeutic 68Ga-DOTATOC PET images provide the best predictive factors of tumor-absorbed doses related to 177Lu-DOTATATE PRRT in refractory meningioma. • A 1.7-fold meninges SUVpeak segmentation method used to determine metabolic tumor volume on pretherapeutic 68Ga-DOTATOC PET images of refractory meningioma treated by 177Lu-DOTATATE is as efficient as the currently routine manual segmentation method and limits inter- and intra-observer variabilities. • This semi-automated method for segmentation of refractory meningioma is easily applicable to routine practice and transferrable across PET centers.

Activity recovery for differently shaped objects in quantitative SPECT

Objective.The aim was to theoretically and experimentally investigate recovery in SPECT images with objects of different shapes. Furthermore, the accuracy of volume estimation by thresholding was studied for those shapes.Approach.Nine spheres, nine oblate spheroids, and nine prolate spheroids phantom inserts were used, of which the six smaller spheres were part of the NEMA IEC body phantom and the rest of the inserts were 3D-printed. The inserts were filled with99mTc and177Lu. When filled with99mTc, SPECT images were acquired in a Siemens Symbia Intevo Bold gamma camera and when filled with177Lu in a General Electric NM/CT 870 DR gamma camera. The signal rate per activity (SRPA) was determined for all inserts and represented as a function of the volume-to-surface ratio and of the volume-equivalent radius using VOIs defined according to the sphere dimensions and VOIs defined using thresholding. Experimental values were compared with theoretical curves obtained analytically (spheres) or numerically (spheroids), starting from the convolution of a source distribution with a point-spread function. Validation of the activity estimation strategy was performed using four 3D-printed ellipsoids. Lastly, the threshold values necessary to determine the volume of each insert were obtained.Main results.Results showed that SRPA values for the oblate spheroids diverted from the other inserts, when SRPA were represented as a function of the volume-equivalent radius. However, SRPA values for all inserts followed a similar behaviour when represented as a function of the volume-to-surface ratio. Results for ellipsoids were in agreement with those results. For the three types of inserts the volume could be accurately estimated using a threshold method for volumes larger than 25 ml.Significance.Determination of SRPA independently of lesion or organ shape should decrease uncertainties in estimated activities and thereby, in the long term, be beneficial to patient care.

Description of a different quantification method for amyloid burden (DPDload) and validation of SPECT/CT in cardiac amyloidosis

BACKGROUND

Bone tracers such as ^99mTc-DPD have shown high sensitivity and specificity in the non-invasive diagnosis of transthyretin cardiac amyloidosis (ATTR-CA). This study aims to validate SPECT/CT and assess the usefulness of uptake quantification (DPDload) in the myocardial tissue as potential information on the amyloid burden.

METHODS

In a retrospective analysis of 46 patients with suspected CA, 23 cases with ATTR-CA had two quantification methods conducted to estimate amyloid burden (DPDload) through planar scintigraphic scans and a SPECT/CT.

RESULTS

SPECT/CT significantly provided an added value in the patient's diagnosis with CA (P<.05). The estimation of the amyloid burden substantiated that the most affected wall of the LV is the interventricular septum in most cases and the existence of a significant relationship between the Perugini score uptake and the DPDload.

CONCLUSIONS

We validate the need for SPECT/CT to complement planar imaging in diagnosing ATTR-CA. For its part, quantifying the amyloid load continues to be a complex area of research. It requires further studies with a larger number of patients to validate a standardized method of amyloid load quantification, both for diagnosis and treatment monitoring.

Relationships between uptake of [68Ga]Ga-DOTA-TATE and absorbed dose in [177Lu]Lu-DOTA-TATE therapy

BACKGROUND

Somatostatin receptor ⁶⁸Ga PET imaging is standard for evaluation of a patient's suitability for ¹⁷⁷Lu peptide receptor radionuclide therapy of neuroendocrine tumours (NETs). The ⁶⁸Ga PET serves to ensure sufficient somatostatin receptor expression, commonly evaluated qualitatively. The aim of this study is to investigate the quantitative relationships between uptake in ⁶⁸Ga PET and absorbed doses in ¹⁷⁷Lu therapy.

METHOD

Eighteen patients underwent [⁶⁸Ga]Ga-DOTA-TATE PET imaging within 20 weeks prior to their first cycle of [¹⁷⁷Lu]Lu-DOTA-TATE. Absorbed doses for therapy were estimated for tumours, kidney, spleen, and normal liver parenchyma using a hybrid SPECT/CT-planar method. Gallium-68 activity concentrations were retrieved from PET images and also used to calculate SUVs and normalized SUVs, using blood and tissue for normalization. The ⁶⁸Ga activity concentrations per injected activity, SUVs, and normalized SUVs were compared with ¹⁷⁷Lu activity concentrations 1 d post-injection and ¹⁷⁷Lu absorbed doses. For tumours, for which there was a variable number per patient, both inter- and intra-patient correlations were analysed. Furthermore, the prediction of ¹⁷⁷Lu tumour absorbed doses based on a combination of tumour-specific ⁶⁸Ga activity concentrations and group-based estimates of the effective half-lives for grade 1 and 2 NETs was explored.

RESULTS

For normal organs, only spleen showed a significant correlation between the ⁶⁸Ga activity concentration and ¹⁷⁷Lu absorbed dose (r = 0.6). For tumours, significant, but moderate, correlations were obtained, with respect to both inter-patient (r = 0.7) and intra-patient (r = 0.45) analyses. The correlations to absorbed doses did not improve when using ⁶⁸Ga SUVs or normalized SUVs. The relationship between activity uptakes for ⁶⁸Ga PET and ¹⁷⁷Lu SPECT was stronger, with correlation coefficients r = 0.8 for both inter- and intra-patient analyses. The ¹⁷⁷Lu absorbed dose to tumour could be predicted from the ⁶⁸Ga activity concentrations with a 95% coverage interval of - 65% to 248%.

CONCLUSIONS

On a group level, a high uptake of [⁶⁸Ga]Ga-DOTA-TATE is associated with high absorbed doses at ¹⁷⁷Lu-DOTA-TATE therapy, but the relationship has a limited potential with respect to individual absorbed dose planning. Using SUV or SUV normalized to reference tissues do not improve correlations compared with using activity concentration per injected activity.

EANM dosimetry committee recommendations for dosimetry of 177Lu-labelled somatostatin-receptor- and PSMA-targeting ligands

The purpose of the EANM Dosimetry Committee is to provide recommendations and guidance to scientists and clinicians on patient-specific dosimetry. Radiopharmaceuticals labelled with lutetium-177 (177Lu) are increasingly used for therapeutic applications, in particular for the treatment of metastatic neuroendocrine tumours using ligands for somatostatin receptors and prostate adenocarcinoma with small-molecule PSMA-targeting ligands. This paper provides an overview of reported dosimetry data for these therapies and summarises current knowledge about radiation-induced side effects on normal tissues and dose-effect relationships for tumours. Dosimetry methods and data are summarised for kidneys, bone marrow, salivary glands, lacrimal glands, pituitary glands, tumours, and the skin in case of radiopharmaceutical extravasation. Where applicable, taking into account the present status of the field and recent evidence in the literature, guidance is provided. The purpose of these recommendations is to encourage the practice of patient-specific dosimetry in therapy with 177Lu-labelled compounds. The proposed methods should be within the scope of centres offering therapy with 177Lu-labelled ligands for somatostatin receptors or small-molecule PSMA.

Dosimetric quantities of neuroendocrine tumors over treatment cycles with 177Lu-DOTA-TATE

Tumor dosimetry was performed for ¹⁷⁷Lu-DOTA-TATE with the aims of better understanding i) the range and variation of the tumor absorbed doses (ADs), ii) how different dosimetric quantities evolve over the treatment cycles, and iii) whether this evolution differs depending on the tumor grade. Such information is important for radiobiological interpretation and may inform the design of alternative administration schemes. Methods: Data come from 41 patients with neuroendocrine tumors (NETs) of grade 1 (n = 23) or 2 (n = 18), that had received between 2 and 9 treatment cycles. Dosimetry was performed for 182 individual lesions, giving in total 880 individual AD assessments across all cycles. Hybrid planar-SPECT/CT imaging was used, including quantitative SPECT reconstruction, voxel-based absorbed-dose-rate calculation, semi-automatic image segmentation, and partial-volume correction. Linear mixed-effect models were used to analyze changes over cycles in tumor ADs, absorbed-dose rates and activity concentrations at day-1, effective half-times, and tumor volumes. Tumors smaller than 8 ml were excluded from analyses. Results: Tumor ADs ranged between 2 and 77 Gy per cycle. On average the AD decreased over the cycles, with significantly different rates (P < 0.05) for grade 1 and 2 NETs of 6% and 14% per cycle, respectively. The absorbed-dose rates and activity concentrations at day-1 decreased by similar amounts. The effective half-times were less variable but shorter for grade 2 than grade 1 (P < 0.001). For grade 2 NETS the tumor volumes decreased, with a similar tendency in grade 1. Conclusion: The tumor AD, absorbed-dose rate and activity uptake decrease, in parallel with tumor volumes, between ¹⁷⁷Lu-DOTA-TATE treatment cycles, particularly for grade 2 NETs. The effective half-times vary less but are lower for grade 2 than grade 1 NETs. These results may indicate the development of radiation-induced fibrosis and could have implications for the design of future treatment and dosimetry protocols.

Learning fuzzy clustering for SPECT/CT segmentation via convolutional neural networks

PURPOSE

Quantitative bone single-photon emission computed tomography (QBSPECT) has the potential to provide a better quantitative assessment of bone metastasis than planar bone scintigraphy due to its ability to better quantify activity in overlapping structures. An important element of assessing the response of bone metastasis is accurate image segmentation. However, limited by the properties of QBSPECT images, the segmentation of anatomical regions-of-interests (ROIs) still relies heavily on the manual delineation by experts. This work proposes a fast and robust automated segmentation method for partitioning a QBSPECT image into lesion, bone, and background.

METHODS

We present a new unsupervised segmentation loss function and its semi- and supervised variants for training a convolutional neural network (ConvNet). The loss functions were developed based on the objective function of the classical Fuzzy C-means (FCM) algorithm. The first proposed loss function can be computed within the input image itself without any ground truth labels, and is thus unsupervised; the proposed supervised loss function follows the traditional paradigm of the deep learning-based segmentation methods and leverages ground truth labels during training. The last loss function is a combination of the first and the second and includes a weighting parameter, which enables semi-supervised segmentation using deep learning neural network.

EXPERIMENTS AND RESULTS

We conducted a comprehensive study to compare our proposed methods with ConvNets trained using supervised, cross-entropy and Dice loss functions, and conventional clustering methods. The Dice similarity coefficient (DSC) and several other metrics were used as figures of merit as applied to the task of delineating lesion and bone in both simulated and clinical SPECT/CT images. We experimentally demonstrated that the proposed methods yielded good segmentation results on a clinical dataset even though the training was done using realistic simulated images. On simulated SPECT/CT, the proposed unsupervised model's accuracy was greater than the conventional clustering methods while reducing computation time by 200-fold. For the clinical QBSPECT/CT, the proposed semi-supervised ConvNet model, trained using simulated images, produced DSCs of 0.75 and 0.74 for lesion and bone segmentation in SPECT, and a DSC of 0.79 bone segmentation of CT images. These DSCs were larger than that for standard segmentation loss functions by > 0.4 for SPECT segmentation, and > 0.07 for CT segmentation with P-values < 0.001 from a paired t-test.

CONCLUSIONS

A ConvNet-based image segmentation method that uses novel loss functions was developed and evaluated. The method can operate in unsupervised, semi-supervised, or fully-supervised modes depending on the availability of annotated training data. The results demonstrated that the proposed method provides fast and robust lesion and bone segmentation for QBSPECT/CT. The method can potentially be applied to other medical image segmentation applications.

Clinical Evaluation of a Three-Dimensional Internal Dosimetry Technique for Liver Radioembolization with 90Y Microspheres Using Dose Voxel Kernels

Background: The purpose of this study was to develop a rapid, reliable, and efficient tool for three-dimensional (3D) dosimetry treatment planning and post-treatment evaluation of liver radioembolization with ⁹⁰Y microspheres, using tissue-specific dose voxel kernels (DVKs) that can be used in everyday clinical practice. Materials and Methods: Two tissue-specific DVKs for ⁹⁰Y were calculated through Monte Carlo (MC) simulations. DVKs for the liver and lungs were generated, and the dose distribution was compared with direct MC simulations. A method was developed to produce a 3D dose map by convolving the calculated DVKs with the activity biodistribution derived from clinical single-photon emission computed tomography (SPECT) or positron emission tomography (PET) images. Image registration for the SPECT or PET images with the corresponding computed tomography scans was performed before dosimetry calculation. The authors first compared the DVK convolution dosimetry with a direct full MC simulation on an XCAT anthropomorphic phantom. They then tested it in 25 individual clinical cases of patients who underwent ⁹⁰Y therapy. All MC simulations were carried out using the GATE MC toolkit. Results: Comparison of the measured absorbed dose using tissue-specific DVKs and direct MC simulation on 25 patients revealed a mean difference of 1.07% ± 1.43% for the liver and 1.03% ± 1.21% for the tumor tissue, respectively. The largest difference between DVK convolution and full MC dosimetry was observed for the lung tissue (10.16% ± 1.20%). The DVK statistical uncertainty was <0.75% for both media. Conclusions: This semiautomatic algorithm is capable of performing rapid, accurate, and efficient 3D dosimetry. The proposed method considers tissue and activity heterogeneity using tissue-specific DVKs. Furthermore, this method provides results in <1 min, making it suitable for everyday clinical practice.

Clinical implementation of PLANET® Dose for dosimetric assessment after [177Lu]Lu-DOTA-TATE: comparison with Dosimetry Toolkit® and OLINDA/EXM® V1.0

Background

The aim of this study was to compare a commercial dosimetry workstation (PLANET® Dose) and the dosimetry approach (GE Dosimetry Toolkit® and OLINDA/EXM® V1.0) currently used in our department for quantification of the absorbed dose (AD) to organs at risk after peptide receptor radionuclide therapy with [¹⁷⁷Lu]Lu-DOTA-TATE.

Methods

An evaluation on phantom was performed to determine the SPECT calibration factor variations over time and to compare the Time Integrated Activity Coefficients (TIACs) obtained with the two approaches. Then, dosimetry was carried out with the two tools in 21 patients with neuroendocrine tumours after the first and second injection of 7.2 ± 0.2 GBq of [¹⁷⁷Lu]Lu-DOTA-TATE (40 dosimetry analyses with each software). SPECT/CT images were acquired at 4 h, 24 h, 72 h and 192 h post-injection and were reconstructed using the Xeleris software (General Electric). The liver, spleen and kidneys masses and TIACs were determined using Dosimetry Toolkit® (DTK) and PLANET® Dose. The ADs were calculated using OLINDA/EXM® V1.0 and the Local Deposition Method (LDM) or Dose voxel-Kernel convolution (DK) on PLANET® Dose.

Results

With the phantom, the 3D calibration factors showed a slight variation (0.8% and 3.3%) over time, and TIACs of 225.19 h and 217.52 h were obtained with DTK and PLANET® Dose, respectively. In patients, the root mean square deviation value was 8.9% for the organ masses, 8.1% for the TIACs, and 9.1% and 7.8% for the ADs calculated with LDM and DK, respectively. The Lin’s concordance correlation coefficient was 0.99 and the Bland–Altman plot analysis estimated that the AD value difference between methods ranged from − 0.75 to 0.49 Gy, from − 0.20 to 0.64 Gy, and from − 0.43 to 1.03 Gy for 95% of the 40 liver, kidneys and spleen dosimetry analyses. The dosimetry method had a minor influence on AD differences compared with the image registration and organ segmentation steps.

Conclusions

The ADs to organs at risk obtained with the new workstation PLANET® Dose are concordant with those calculated with the currently used software and in agreement with the literature. These results validate the use of PLANET® Dose in clinical routine for patient dosimetry after targeted radiotherapy with [¹⁷⁷Lu]Lu-DOTA-TATE.

Cardiac transthyretin amyloidosis 99mTc-DPD SPECT correlates with strain echocardiography and biomarkers

Purpose

Hereditary transthyretin-amyloid amyloidosis (ATTRv) is an underdiagnosed condition commonly manifesting as congestive heart failure. Recently, scintigraphy utilizing DPD as a tracer was shown to identify ATTRv and wild-type ATTR cardiomyopathy. The aim of this study was to determine the value of quantified scintigraphy utilizing ^99mTc-3,3-diphosphono-1,2-propanodicarboxylic acid (DPD) single-photon emission computed tomography (SPECT)/CT, and to correlate its uptake with well-established cardiac functional parameters.

Methods

Forty-eight patients with genetically verified ATTRv type-A fibril composition, positive ^99mTc-DPD SPECT/CT, were retrospectively analyzed. Manual mapping of volumes of interest (VOIs) on DPD SPECT/CT examinations was used to quantify heart uptake. DPD mean and maximum uptake together with a calculated DPD-based amyloid burden (DPD_load) was correlated with echocardiographic strain values and cardiac biomarkers.

Results

Statistically significant correlations were seen in VOIs between DPD uptakes and the corresponding echocardiographic strain values. Furthermore, DPD_load had a strong correlation with echocardiographic strain parameters and also correlated with biomarkers troponin T and logarithmic NT-ProBNP.

Conclusions

In patients with ATTRv cardiomyopathy, DPD SPECT/CT measures the amyloid distribution and provides information on cardiac amyloid load. DPD amyloid load correlates with functional cardiac parameters.

Feasibility and limitations of quantitative SPECT for 223Ra

The aim of this paper is to investigate the feasibility and limitations of activity-concentration estimation for ²²³Ra using SPECT. Phantom measurements are performed using spheres (volumes 5.5 mL to 26.4 mL, concentrations 1.6 kBq mL^-1 to 4.5 kBq mL^-1). Furthermore, SPECT projections are simulated using the SIMIND Monte Carlo program for two geometries, one similar to the physical phantom and the other being an anthropomorphic phantom with added lesions (volumes 34 mL to 100 mL, concentrations 0.5 kBq mL^-1 to 4 kBq mL^-1). Medium-energy and high-energy collimators, 60 projections with 55 s per projection and a 20% energy window at 82 keV are employed. For the Monte Carlo simulated images, Poisson-distributed noise is added in ten noise realizations. Reconstruction is performed (OS-EM, 40 iterations, 6 subsets) employing compensation for attenuation, scatter, and collimator-detector response. The estimated concentrations in the anthropomorphic phantom are also corrected using recovery coefficients. Errors for the largest sphere in the physical phantom range from -25% to -34% for the medium-energy collimator and larger deviations for smaller spheres. Corresponding results for the high-energy collimator are -15% to -31%. The corresponding Monte Carlo simulations show standard deviations of a few percentage points. For the anthropomorphic phantom, before application of recovery coefficients the bias ranges from -16% to -46% (medium-energy collimator) and -10% to -28% (high-energy collimator), with standard deviations of 2% to 14% and 1% to 16%. After the application of recovery coefficients, the biases range from -3% to -35% (medium energy collimator) and from 0% to -18%. The errors decrease with increasing concentrations. Activity-concentration estimation of ²²³Ra with SPECT is feasible, but problems with repeatability need to be further studied.

Personalized kidney dosimetry in 177Lu-octreotate treatment of neuroendocrine tumours: a comparison of kidney dosimetry estimates based on a whole organ and small volume segmentations

Peptide receptor radionuclide therapy (PRRT) with ¹⁷⁷Lu- radiolabeled octreotate is an effective treatment method for inoperable neuroendocrine tumours (NETs). There is growing evidence that estimates of the organ-at-risks (OARs) doses are necessary for the optimization of personalized PRRT (P-PRRT). Dosimetry, however, requires a complicated and time-consuming procedure, which hinders its implementation in the clinic. The aim of this study is to develop a practical and automatic technique to simplify personalized dosimetry of kidney, the major OAR in ¹⁷⁷Lu P-PRRT. The data from 30 NETs patients undergoing 44 personalized ¹⁷⁷Lu-DOTA-TATE therapy cycles were analyzed. To determine the biokinetics of the radiopharmaceutical in the kidneys, for each patient three SPECT/CT scans were acquired, at about 4 h, 24 h and 70 h after injection. The kidneys doses were evaluated using three different approaches: (1) a traditional approach based on whole kidney (WK) segmentation; (2) a small volume (SV) manual approach (M-SV) with observer-defined SV location; and (3) a software based SV-approach that automatically defines SV location (A-SV). Four different methods of automatic SV location selections were investigated. The SV kidney doses estimated using M-SV and A-SV approaches was evaluated and the accuracy of these two approaches were compared to the WK dosimetry. The kidney bio-kinetics, in terms of effective half-lives, obtained from both of the A-SV and M-SV approaches agreed to within 10% with those obtained from the WK segmentation. The average ratios of SV doses to WK doses were mostly about 1.8  ±  0.2 for both A-SV and M-SV approaches. The linear correlation coefficients between SV doses (both A-SV and M-SV) and WK doses were up to 0.9 with p   <  0.001. The differences between A-SV and M-SV were minor. By comparing different methods of SV location selections, independently selecting SV in images from each of the acquisitions was proved the most appropriate and accurate approach. An automatic, observer-independent method for selecting the location of the small volume in kidneys was developed. The accuracy of this dose estimation approach has been demonstrated by comparing it with the manual SV dosimetry, as well as the WK dosimetry. The proposed automatic approach can potentially be considered as a practical and simple method for dose estimation in the future clinical studies.

A method for tumor dosimetry based on hybrid planar-SPECT/CT images and semiautomatic segmentation

PURPOSE

A hybrid planar-SPECT/CT method for tumor dosimetry in 177 Lu-DOTATATE therapy, applicable to datasets consisting of multiple conjugate-view images and one SPECT/CT, is developed and evaluated.

METHODS

The imaging protocol includes conjugate-view imaging at 1, 24, 96, and 168 h post infusion (p.i.) and a SPECT/CT acquisition 24 h p.i. The dosimetry method uses the planar images to estimate the shape of the time-activity concentration curve, which is then rescaled to absolute units using the SPECT-derived activity concentration. The resulting time-integrated activity concentration coefficient (TIACC) is used to calculate the tumor-absorbed dose. Semiautomatic segmentation techniques are applied for tumor delineation in both planar and SPECT images, where the planar image segmentation is accomplished using an active-rays-based technique. The selection of tumors is done by visual inspection of planar and SPECT images and applying a set of criteria concerning the tumor visibility and possible interference from superimposed activity uptakes in the planar images. Five different strategies for determining values from planar regions of interest (ROIs), based on entire or partial ROIs, and with and without background correction, are evaluated. Evaluation is performed against a SPECT/CT-based method on data from six patients where sequential conjugate-view and SPECT/CT imaging have been performed in parallel and against ground truths in Monte Carlo simulated images. The patient data are also used to evaluate the interoperator variability and to assess the validity of the developed criteria for tumor selection.

RESULTS

For patient images, the hybrid method produces TIACCs that are on average 6% below those of the SPECT/CT only method, with standard deviations for the relative TIACC differences of 8%-11%. Simulations show that the hybrid and SPECT-based methods estimate the TIACCs to within approximately 10% for tumors larger than around 10 ml, while for smaller tumors, all methods underestimate the TIACCs due to underestimations of the activity concentrations in the SPECT images. The planar image segmentation has a low operator dependence, with a median Dice similarity coefficient of 0.97 between operators. The adopted criteria for tumor selection manage to discriminate the tumors for which the absorbed-dose deviations between the hybrid and SPECT methods are the highest.

CONCLUSIONS

The hybrid method is found suitable for studies of tumor-absorbed doses in radionuclide therapy, provided that selection criteria regarding the visibility and overlapping activities in the planar images are applied.

The Relevance of Dosimetry in Precision Medicine

The aim of this review is to provide an overview of the most recent technologic developments in state-of-the-art equipment and tools for dosimetry in radionuclide therapies. This includes, but is not restricted to, calibration methods for imaging systems. In addition, a summary of new developments that consider the influence of small-scale dosimetry and of biologic effects on radionuclide therapies is given. Finally, the current limitations of patient-specific dosimetry such as bone-marrow dosimetry or dosimetry of α-emitters are discussed.

Optimizing Image Quantification for 177Lu SPECT/CT Based on a 3D Printed 2-Compartment Kidney Phantom

The aim of this work was to find an optimal setup for activity determination of ¹⁷⁷Lu-based SPECT/CT imaging reconstructed with 2 commercially available methods (xSPECT Quant and Flash3D). For this purpose, 3-dimensional (3D)-printed phantoms of different geometries were manufactured, different partial-volume correction (PVC) methods were applied, and the accuracy of the activity determination was evaluated. Methods: A 2-compartment kidney phantom (70% cortical and 30% medullary compartment), a sphere, and an ellipsoid of equal volumes were 3D printed, filled with ¹⁷⁷Lu, and scanned with a SPECT/CT system. Reconstructions were performed with xSPECT and Flash3D. Different PVC methods were applied to find an optimal quantification setup: method 1 was a geometry-specific recovery coefficient based on the 3D printing model, method 2 was a geometry-specific recovery coefficient based on the low-dose CT scan, method 3 was an enlarged volume of interest including spilled-out counts, method 4 was activity concentration in the peak milliliter applied to the entire CT-based volume, and method 5 was a fixed threshold of 42% of the maximum in a large volume containing the object of interest. Additionally, the influence of postreconstruction gaussian filtering was investigated. Results: Although the recovery coefficients of sphere and ellipsoid differed by only 0.7%, a difference of 31.7% was observed between the sphere and the renal cortex phantoms. Without postfiltering, the model-based recovery coefficients (methods 1 and 2) resulted in the best accuracies (xSPECT, 1.5%; Flash3D, 10.3%), followed by the enlarged volume (method 3) (xSPECT, 8.5%; Flash3D, 13.0%). The peak-milliliter method (method 4) showed large errors only for sphere and ellipsoid (xSPECT, 23.4%; Flash3D, 21.6%). Applying a 42% threshold (method 5) led to the largest quantification errors (xSPECT, 32.3%; Flash3D, 46.7%). After postfiltering, a general increase in the errors was observed. Conclusion: In this work, 3D printing was used as a prototyping technique for a geometry-specific investigation of SPECT/CT reconstruction parameters and PVC methods. The optimal setup for activity determination was found to be an unsmoothed SPECT/CT reconstruction in combination with a recovery coefficient based on the low-dose CT. The difference between spheric and renal recovery coefficients suggests that the typically applied volume-dependent but only sphere-based recovery coefficient lookup tables should be replaced by a more geometry-specific alternative.