Fit of biokinetic data in molecular radiotherapy: a machine learning approach.

BACKGROUND: In literature are reported different analytical methods (AM) to choose the proper fit model and to fit data of the time-activity curve (TAC). On the other hand, Machine Learning algorithms (ML) are increasingly used for both classification and regression tasks. The aim of this work was to investigate the possibility of employing ML both to classify the most appropriate fit model and to predict the area under the curve (τ). METHODS: Two different ML systems have been developed for classifying the fit model and to predict the biokinetic parameters. The two systems were trained and tested with synthetic TACs simulating a whole-body Fraction Injected Activity for patients affected by metastatic Differentiated Thyroid Carcinoma, administered with [131I]I-NaI. Test performances, defined as classification accuracy (CA) and percentage difference between the actual and the estimated area under the curve (Δτ), were compared with those obtained using AM varying the number of points (N) of the TACs. A comparison between AM and ML were performed using data of 20 real patients. RESULTS: As N varies, CA remains constant for ML (about 98%), while it improves for F-test (from 62 to 92%) and AICc (from 50 to 92%), as N increases. With AM, [Formula: see text] can reach down to - 67%, while using ML [Formula: see text] ranges within ± 25%. Using real TACs, there is a good agreement between τ obtained with ML system and AM. CONCLUSIONS: The employing of ML systems may be feasible, having both a better classification and a better estimation of biokinetic parameters.

Evolution of Portable Sensors for In-Vivo Dose and Time-Activity Curve Monitoring as Tools for Personalized Dosimetry in Molecular Radiotherapy.

Treatment personalization in Molecular Radiotherapy (MRT) relies on pre- and post-treatment SPECT/ PET-based images and measurements to obtain a patient-specific absorbed dose-rate distribution map and its evolution over time. Unfortunately, the number of time points that are available per patient to investigate individual pharmacokinetics is often reduced by limited patient compliance or SPECT or PET/CT scanner availability for dosimetry in busy departments. The adoption of portable sensors for in-vivo dose monitoring during the entire treatment could improve the assessment of individual biokinetics in MRT and, thus, the treatment personalization. The evolution of portable devices, non-SPECT/PET-based options, already used for monitoring radionuclide activity transit and accumulation during therapy with radionuclides (i.e., MRT or brachytherapy), is presented to identify valuable ones, which combined with conventional nuclear medicine imaging systems could be effective in MRT. External probes, integration dosimeters and active detecting systems were included in the study. The devices and their technology, the range of applications, the features and limitations are discussed. Our overview of the available technologies encourages research and development of portable devices and dedicated algorithms for MRT patient-specific biokinetics study. This would represent a crucial advancement towards personalized treatment in MRT.

Localization of anatomical changes in patients during proton therapy with in-beam PET monitoring: A voxel-based morphometry approach exploiting Monte Carlo simulations.

PURPOSE: In-beam positron emission tomography (PET) is one of the modalities that can be used for in vivo noninvasive treatment monitoring in proton therapy. Although PET monitoring has been frequently applied for this purpose, there is still no straightforward method to translate the information obtained from the PET images into easy-to-interpret information for clinical personnel. The purpose of this work is to propose a statistical method for analyzing in-beam PET monitoring images that can be used to locate, quantify, and visualize regions with possible morphological changes occurring over the course of treatment. METHODS: We selected a patient treated for squamous cell carcinoma (SCC) with proton therapy, to perform multiple Monte Carlo (MC) simulations of the expected PET signal at the start of treatment, and to study how the PET signal may change along the treatment course due to morphological changes. We performed voxel-wise two-tailed statistical tests of the simulated PET images, resembling the voxel-based morphometry (VBM) method commonly used in neuroimaging data analysis, to locate regions with significant morphological changes and to quantify the change. RESULTS: The VBM resembling method has been successfully applied to the simulated in-beam PET images, despite the fact that such images suffer from image artifacts and limited statistics. Three dimensional probability maps were obtained, that allowed to identify interfractional morphological changes and to visualize them superimposed on the computed tomography (CT) scan. In particular, the characteristic color patterns resulting from the two-tailed statistical tests lend themselves to trigger alarms in case of morphological changes along the course of treatment. CONCLUSIONS: The statistical method presented in this work is a promising method to apply to PET monitoring data to reveal interfractional morphological changes in patients, occurring over the course of treatment. Based on simulated in-beam PET treatment monitoring images, we showed that with our method it was possible to correctly identify the regions that changed. Moreover we could quantify the changes, and visualize them superimposed on the CT scan. The proposed method can possibly help clinical personnel in the replanning procedure in adaptive proton therapy treatments.

Technical note: A wearable radiation measurement system for collection of patient-specific time-activity data in radiopharmaceutical therapy: system design and Monte Carlo simulation results.

PURPOSE: A high level of personalization in Molecular Radiotherapy (MRT) could bring advantages in terms of treatment effectiveness and toxicity reduction. Individual organ-level dosimetry is crucial to describe the radiopharmaceutical biodistribution expressed by the patient, to estimate absorbed doses to normal organs and target tissue(s). This paper presents a proof-of-concept Monte Carlo simulation study of "WIDMApp" (Wearable Individual Dose Monitoring Apparatus), a multi-channel radiation detector and data processing system for in vivo patient measurement and collection of radiopharmaceutical biokinetic data (i.e., time-activity data). Potentially, such a system can increase the amount of such data that can be collected while reducing the need to derive it via nuclear medicine imaging. METHODS: a male anthropomorphic MIRD phantom was used to simulate photons (i.e., gamma-rays) propagation in a patient undergoing a 131 I thyroid treatment. The administered activity was set to the amount usually administered for the treatment of differentiated carcinoma while its initial distribution in different organs was assigned following the ICRP indications for the 131 I biokinetics. Using this information, the simulation computes the Time-dependent Counts Curves (TCCs) that would have been measured by seven WIDMApp-like sensors placed and oriented to face each one of five emitting organs plus two thyroid lobes. A deconvolution algorithm was then applied on this simulated data set to reconstruct the Time-Activity Curve (TAC) of each organ. Deviations of the reconstructed TACs parameters from values used to generate them were studied as a function of the deconvolution algorithm initialization parameters and assuming non-Poisson fluctuation of the TCCs data points. RESULTS: This study demonstrates that it is possible, at least in the simple simulated scenario, to reconstruct the organ cumulated activity by measuring the time dependence of counts recorded by several detectors placed at selected positions on the patient's body. The ability to perform in vivo sampling more frequently than conventional biokinetic studies increases the number of time points and therefore the accuracy in TAC estimates. In this study, an accuracy on cumulated activity of 5% is obtained even with a 20% error on the TCC data points and a 50% error on the initial guess on the parameters of the deconvolution algorithm. CONCLUSIONS: the WIDMApp approach could provide an effective tool to characterize more accurately the radiopharmaceutical biokinetics in MRT patients, reducing the need of resources of nuclear medicine departments, such as technologist and scanner time, to perform individualized biokinetics studies. The relatively simple hardware for the approach proposed would allow its application to large numbers of patients. The results obtained justify development of an actual prototype system to characterize this technique under realistic conditions.

Monitoring Carbon Ion Beams Transverse Position Detecting Charged Secondary Fragments: Results From Patient Treatment Performed at CNAO.

Particle therapy in which deep seated tumours are treated using 12C ions (Carbon Ions RadioTherapy or CIRT) exploits the high conformity in the dose release, the high relative biological effectiveness and low oxygen enhancement ratio of such projectiles. The advantages of CIRT are driving a rapid increase in the number of centres that are trying to implement such technique. To fully profit from the ballistic precision achievable in delivering the dose to the target volume an online range verification system would be needed, but currently missing. The 12C ions beams range could only be monitored by looking at the secondary radiation emitted by the primary beam interaction with the patient tissues and no technical solution capable of the needed precision has been adopted in the clinical centres yet. The detection of charged secondary fragments, mainly protons, emitted by the patient is a promising approach, and is currently being explored in clinical trials at CNAO. Charged particles are easy to detect and can be back-tracked to the emission point with high efficiency in an almost background-free environment. These fragments are the product of projectiles fragmentation, and are hence mainly produced along the beam path inside the patient. This experimental signature can be used to monitor the beam position in the plane orthogonal to its flight direction, providing an online feedback to the beam transverse position monitor chambers used in the clinical centres. This information could be used to cross-check, validate and calibrate, whenever needed, the information provided by the ion chambers already implemented in most clinical centres as beam control detectors. In this paper we study the feasibility of such strategy in the clinical routine, analysing the data collected during the clinical trial performed at the CNAO facility on patients treated using 12C ions and monitored using the Dose Profiler (DP) detector developed within the INSIDE project. On the basis of the data collected monitoring three patients, the technique potential and limitations will be discussed.

First Ex Vivo Results of ß--Radioguided Surgery in Small Intestine Neuroendocrine Tumors with 90Y-DOTATOC.

Background: In neuroendocrine tumor (NET), complete surgery could better the prognosis. Radioguided surgery (RGS) with ß--radioisotopes is a novel approach focused on developing a new probe that, detecting electrons and operating with low background, provides a clearer delineation of the lesions with low radiation exposition for surgeons. As a first step to validate this procedure, ex vivo specimens of tumors expressing somatostatin receptors, as small intestine neuroendocrine tumor (SI-NET), were tested. Materials and Methods: SI-NET presents a high uptake of a beta-emitting radiotracer, 90Y-DOTATOC. Five SI-NET patients were enrolled after performing a 68Ga-DOTATOC positron emission tomography/computed tomography (CT) and a CT enterography; 24 h before surgery, they received 5 mCi of 90Y-DOTATOC. Results: Surgery was performed as routine. Tumors and surrounding tissue were sectioned in different samples and examined ex vivo with the beta-detecting probe. All the tumor samples showed high counts of radioactivity that was up to a factor of 18 times higher than the corresponding cutoff value, with a sensitivity of 96% and a specificity of 100%. Conclusions: These first ex vivo RGS tests showed that this probe can discriminate very effectively between tumor and healthy tissues by the administration of low activities of 90Y-DOTATOC, allowing more precise surgery.

A DROP-IN beta probe for robot-assisted 68Ga-PSMA radioguided surgery: first ex vivo technology evaluation using prostate cancer specimens.

BACKGROUND: Recently, a flexible DROP-IN gamma-probe was introduced for robot-assisted radioguided surgery, using traditional low-energy SPECT-isotopes. In parallel, a novel approach to achieve sensitive radioguidance using beta-emitting PET isotopes has been proposed. Integration of these two concepts would allow to exploit the use of PET tracers during robot-assisted tumor-receptor-targeted. In this study, we have engineered and validated the performance of a novel DROP-IN beta particle (DROP-INß) detector. METHODS: Seven prostate cancer patients with PSMA-PET positive tumors received an additional intraoperative injection of ~ 70 MBq 68Ga-PSMA-11, followed by robot-assisted prostatectomy and extended pelvic lymph node dissection. The surgical specimens from these procedures were used to validate the performance of our DROP-INß probe prototype, which merged a scintillating detector with a housing optimized for a 12-mm trocar and prograsp instruments. RESULTS: After optimization of the detector and probe housing via Monte Carlo simulations, the resulting DROP-INß probe prototype was tested in a robotic setting. In the ex vivo setting, the probe-positioned by the robot-was able to identify 68Ga-PSMA-11 containing hot-spots in the surgical specimens: signal-to-background (S/B) was > 5 when pathology confirmed that the tumor was located < 1 mm below the specimen surface. 68Ga-PSMA-11 containing (and PET positive) lymph nodes, as found in two patients, were also confirmed with the DROP-INß probe (S/B > 3). The rotational freedom of the DROP-IN design and the ability to manipulate the probe with the prograsp tool allowed the surgeon to perform autonomous beta-tracing. CONCLUSIONS: This study demonstrates the feasibility of beta-radioguided surgery in a robotic context by means of a DROP-INß detector. When translated to an in vivo setting in the future, this technique could provide a valuable tool in detecting tumor remnants on the prostate surface and in confirmation of PSMA-PET positive lymph nodes.

Radioguided surgery with ß- radiation in pancreatic Neuroendocrine Tumors: a feasibility study.

The possibility to use ß- decaying isotopes for radioguided surgery (RGS) has been recently proposed, and first promising tests on ex-vivo samples of Meningioma and intestinal Neuroendocrine Tumor (NET) have been published. This paper reports a study of the uptake of 68Ga-DOTATOC in pancreatic NETs (pNETs) in order to assess the feasibility of a new RGS approach using 90Y-DOTATOC. Tumor and healthy pancreas uptakes were estimated from 68Ga-DOTATOC PET/CT scans of 30 patients with pNETs. From the obtained SUVs (Standardised Uptake Value) and TNRs (Tumor Non tumor Ratio), an analysis algorithm relying on a Monte Carlo simulation of the detector has been applied to evaluate the performances of the proposed technique. Almost all considered patients resulted to be compatible with the application of ß--RGS assuming to administer 1.5 MBq/kg of activity of 90Y-DOTATOC 24 h before surgery, and a sampling time of few seconds. In just 2 cases the technique would have required a mildly increased amount of activity or of sampling time. Despite a high physiological uptake of 68Ga-DOTATOC in the healthy pancreas, the proposed RGS technique promises to be effective. This approach allows RGS to find application also in pancreatic diseases, where traditional techniques are not viable.

Radioguided surgery with ß radiation: a novel application with Ga68.

Radio Guided Surgery is a technique helping the surgeon in the resection of tumors: a radiolabeled tracer is administered to the patient before surgery and then the surgeon evaluates the completeness of the resection with a handheld detector sensitive to emitted radiation. Established methods rely on γ emitting tracers coupled with γ detecting probes. The efficacy of this technique is however hindered by the high penetration of γ radiation, limiting its applicability to low background conditions. To overtake such limitations, a novel approach to RGS has been proposed, relying on ß- emitting isotopes together with a dedicated ß probe. This technique has been proved to be effective in first ex-vivo trials. We discuss in this paper the possibility to extend its application cases to 68Ga, a ß+ emitting isotope widely used today in nuclear medicine. To this aim, a retrospective study on 45 prostatic cancer patients was performed, analysing their 68Ga-PSMA PET images to asses if the molecule uptake is enough to apply this technique. Despite the expected variability both in terms of SUV (median 4.1, IQR 3.0-6.1) and TNR (median 9.4, IQR 5.2-14.6), the majority of cases have been found to be compatible with ß-RGS with reasonable injected activity and probing time (5 s).

Mass spectrometry characterization of DOTA-Nimotuzumab conjugate as precursor of an innovative ß- tracer suitable in radio-guided surgery.

The aim of the present work has been the mass spectrometry characterization of the Nimotuzumab (NIM) antibody chemically modified with the bifunctional chelating agent para-S-2-(4-isothiocyanatobenzyl)-1,4,7,10-tetraaza cyclododecanetetraacetic acid (p-SCN-Bn-DOTA). The conjugate, upon labeling with the pure ß--emitter 90Y3+, could represent a promising candidate as radiotracer for an innovative radio-guided surgery (RGS) technique, developed and patented by researchers of our group, which uses a probe system for intraoperative detection of tumor residues exploiting the selective uptake of ß--emitting tracers. The results reported in this study show that multiple DOTA molecules bind to lysine residues of both light and heavy chains of the antibody and, probably, some of them are linked to the variable region of antibody. Moreover, the new mass spectrometric analysis highlights the presence of unreacted NIM in the final product. The information obtained by this work is of fundamental importance in the perspective to utilize this conjugate as a radiocompound after its labeling with 90Y3+ radioisotope. Indeed, the conjugation efficiency and the presence of unreacted NIM affect the specific activity of the final radiotracer which binds specific receptor.

Feasibility of beta-particle radioguided surgery for a variety of "nuclear medicine" radionuclides.

PURPOSE: Beta-particle radioguided tumor resection may potentially overcome the limitations of conventional gamma-ray guided surgery by eliminating, or at least minimizing, the confounding effect of counts contributed by activity in adjacent normal tissues. The current study evaluates the clinical feasibility of this approach for a variety of radionuclides. Nowadays, the only ß- radioisotope suited to radioguided surgery is 90Y. Here, we study the ß- probe prototype capability to different radionuclides chosen among those used in nuclear medicine. METHODS: The counting efficiency of our probe prototype was evaluated for sources of electrons and photons of different energies. Such measurements were used to benchmark the Monte Carlo (MC) simulation of the probe behavior, especially the parameters related to the simulation of the optical photon propagation in the scintillation crystal. Then, the MC simulation was used to derive the signal and the background we would measure from a small tumor embedded in the patient body if one of the selected radionuclides is used. RESULTS: Based on the criterion of detectability of a 0.1 ml tumor for a counting interval of 1 s and an administered activity of 3 MBq/kg, the current probe yields a detectable signal over a wide range of Standard Uptake Values (SUVs) and tumor-to-non-tumor activity-concentration ratios (TNRs) for 31Si, 32P, 97Zr, and 188Re. Although efficient counting of 83Br, 133I, and 153Sm proved somewhat more problematic, the foregoing criterion can be satisfied for these isotopes as well for sufficiently high SUVs and TNRs.

Addendum: Measurement of charged particle yields from PMMA irradiated by a 220 MeV/u 12C beam.

In this paper we report the re-analysis of the data published in (Piersanti et al. 2014) documenting the charged secondary particles production induced by the interaction of a 220 MeV/u 12C ion beam impinging on a polymethyl methacrylate (PMMA) target, measured in 2012 at the GSI facility in Darmstadt (Germany). This re-analysis takes into account the inhomogeneous light response of the LYSO crystal in the experimental setup measured in a subsequent experiment (2014) performed in the Heidelberg Ion- Beam Therapy Center. A better description of the detector and re-calculation of the geometrical efficiencies have been implemented as well, based on an improved approach that accounts also for the energy dependence of the emission spectrum. The new analysis has small effect on the total secondary charged flux, but has an impact on the production yield and emission velocity distributions of the different particle species (protons, deuterons and tritons) at different angles with respect to the beam direction (60° and 90°). All these observables indeed depend on the particle identification algorithms and hence on the LYSO detector energy response. The results of the data re-analysis presented here are intended to supersede and replace the results published in (Piersanti et al. 2014).

Design of a new tracking device for on-line beam range monitor in carbon therapy.

Charged particle therapy is a technique for cancer treatment that exploits hadron beams, mostly protons and carbon ions. A critical issue is the monitoring of the beam range so to check the correct dose deposition to the tumor and surrounding tissues. The design of a new tracking device for beam range real-time monitoring in pencil beam carbon ion therapy is presented. The proposed device tracks secondary charged particles produced by beam interactions in the patient tissue and exploits the correlation of the charged particle emission profile with the spatial dose deposition and the Bragg peak position. The detector, currently under construction, uses the information provided by 12 layers of scintillating fibers followed by a plastic scintillator and a pixelated Lutetium Fine Silicate (LFS) crystal calorimeter. An algorithm to account and correct for emission profile distortion due to charged secondaries absorption inside the patient tissue is also proposed. Finally detector reconstruction efficiency for charged particle emission profile is evaluated using a Monte Carlo simulation considering a quasi-realistic case of a non-homogenous phantom.

Monitoring of Hadrontherapy Treatments by Means of Charged Particle Detection.

The interaction of the incoming beam radiation with the patient body in hadrontherapy treatments produces secondary charged and neutral particles, whose detection can be used for monitoring purposes and to perform an on-line check of beam particle range. In the context of ion-therapy with active scanning, charged particles are potentially attractive since they can be easily tracked with a high efficiency, in presence of a relatively low background contamination. In order to verify the possibility of exploiting this approach for in-beam monitoring in ion-therapy, and to guide the design of specific detectors, both simulations and experimental tests are being performed with ion beams impinging on simple homogeneous tissue-like targets (PMMA). From these studies, a resolution of the order of few millimeters on the single track has been proven to be sufficient to exploit charged particle tracking for monitoring purposes, preserving the precision achievable on longitudinal shape. The results obtained so far show that the measurement of charged particles can be successfully implemented in a technology capable of monitoring both the dose profile and the position of the Bragg peak inside the target and finally lead to the design of a novel profile detector. Crucial aspects to be considered are the detector positioning, to be optimized in order to maximize the available statistics, and the capability of accounting for the multiple scattering interactions undergone by the charged fragments along their exit path from the patient body. The experimental results collected up to now are also valuable for the validation of Monte Carlo simulation software tools and their implementation in Treatment Planning Software packages.

Time Evolution of DOTATOC Uptake in Neuroendocrine Tumors in View of a Possible Application of Radioguided Surgery with ß- Decay.

UNLABELLED: A novel radioguided surgery (RGS) technique exploiting ß- radiation has been proposed. To develop such a technique, a suitable radiotracer able to deliver a ß- emitter to the tumor has to be identified. A first candidate is represented by 90Y-labeled DOTATOC, a compound commonly used today for peptide radioreceptor therapy. The application of this ß- RGS to neuroendocrine tumors (NET) requires study of the uptake of DOTATOC and its time evolution both in tumors and in healthy organs and evaluation of the corresponding performance of the technique. METHODS: Uptake by lesions and healthy organs (kidneys, spleen, liver and healthy muscle) was estimated on 177Lu-DOTATOC SPECT/CT scans of 15 patients affected by NET with different localizations, treated at IRCCS-Arcispedale Santa Maria Nuova, Reggio Emilia, Italy. For each patient, SPECT/CT images, acquired at 0.5, 4, 20, 40, and 70 h after injection, were studied. For each lesion, the tumor-to-nontumor ratio (TNR) with respect to all healthy organs and its time evolution were studied. A subset of patients showing hepatic lesions was selected, and the TNR with respect to the nearby healthy tissue was calculated. By means of a Monte Carlo simulation of the probe for ß- RGS, the activity that is to be administered for a successful detection was estimated lesion-by-lesion. RESULTS: Uptake of DOTATOC on NETs maximized at about 24 h after injection. The cases of hepatic lesions showed a TNR with respect to the tumor margins compatible with the application of ß- RGS. In particular, 0.1-mL residuals are expected to be detectable within 1 s with 5% false-negative and 1% false-positive by administering the patient as little as 1 MBq/kg. CONCLUSION: The balance between tumor uptake and metabolic washout in healthy tissue causes the TNR to increase with time, reaching its maximum after 24 h, and this characteristic can be exploited when a radiotracer with a long half-life, such as 90Y, is used. In particular, if 90Y-DOTATOC is used with liver NET metastases, the proposed RGS technique is believed to be feasible by injecting an activity that is one third of that commonly used for PET imaging.

Toward radioguided surgery with ß- decays: uptake of a somatostatin analogue, DOTATOC, in meningioma and high-grade glioma.

UNLABELLED: A novel radioguided surgery (RGS) technique for cerebral tumors using ß(-) radiation is being developed. Checking for a radiotracer that can deliver a ß(-) emitter to the tumor is a fundamental step in the deployment of such a technique. This paper reports a study of the uptake of (90)Y-DOTATOC in meningiomas and high-grade gliomas (HGGs) and a feasibility study of the RGS technique in these types of tumor. Estimates were performed assuming the use of a ß(-) probe under development with a sensitive area 2.55 mm in radius to detect 0.1-mL residuals. METHODS: Uptake and background from healthy tissues were estimated on (68)Ga-DOTATOC PET scans of 11 meningioma patients and 12 HGG patients. A dedicated statistical analysis of the DICOM images was developed and validated. The feasibility study was performed using full simulation of emission and detection of the radiation, accounting for the measured uptake and background rate. RESULTS: All meningioma patients but one with an atypical extracranial tumor showed high uptake of DOTATOC. In terms of feasibility of the RGS technique, we estimated that by administering a 3 MBq/kg activity of radiotracer, the time needed to detect a 0.1-mL remnant with 5% false-negative and 1% false-positive rates is less than 1 s. Actually, to achieve a detection time of 1 s the required activities to administer were as low as 0.2-0.5 MBq/kg in many patients. In HGGs, the uptake was lower than in meningiomas, but the tumor-to-nontumor ratio was higher than 4, which implies that the tracer can still be effective for RGS. It was estimated that by administering 3 mBq/kg of radiotracer, the time needed to detect a 0.1-mL remnant is less than 6 s, with the exception of the only oligodendroma in the sample. CONCLUSION: Uptake of (90)Y-DOTATOC in meningiomas was high in all studied patients. Uptake in HGGs was significantly worse than in meningiomas but was still acceptable for RGS, particularly if further research and development are done to improve the performance of the ß(-) probe.