A novel adaptive needle insertion sequencing for robotic, single needle MR-guided high-dose-rate prostate brachytherapy

MR-guided high-dose-rate (HDR) brachytherapy has gained increasing interest as a treatment for patients with localized prostate cancer because of the superior value of MRI for tumor and surrounding tissues localization. To enable needle insertion into the prostate with the patient in the MR bore, a single needle MR-compatible robotic system involving needle-by-needle dose delivery has been developed at our institution. Throughout the intervention, dose delivery may be impaired by: (1) sub-optimal needle positioning caused by e.g. needle bending, (2) intra-operative internal organ motion such as prostate rotations or swelling, or intra-procedural rectum or bladder filling. This may result in failure to reach clinical constraints. To assess the first aforementioned challenge, a recent study from our research group demonstrated that the deposited dose may be greatly improved by real-time adaptive planning with feedback on the actual needle positioning. However, the needle insertion sequence is left to the doctor and therefore, this may result in sub-optimal dose delivery. In this manuscript, a new method is proposed to determine and update automatically the needle insertion sequence. This strategy is based on the determination of the most sensitive needle track. The sensitivity of a needle track is defined as its impact on the dose distribution in case of sub-optimal positioning. A stochastic criterion is thus presented to determine each needle track sensitivity based on needle insertion simulations. To assess the proposed sequencing strategy, HDR prostate brachytherapy was simulated on 11 patients with varying number of needle insertions. Sub-optimal needle positioning was simulated at each insertion (modeled by typical random angulation errors). In 91% of the scenarios, the dose distribution improved when the needle was inserted into the most compared to the least sensitive needle track. The computation time for sequencing was less than 6 s per needle track. The proposed needle insertion sequencing can therefore assist in delivering an optimal dose in HDR prostate brachytherapy.

Adaptive planning strategy for high dose rate prostate brachytherapy—a simulation study on needle positioning errors

The development of magnetic resonance (MR) guided high dose rate (HDR) brachytherapy for prostate cancer has gained increasing interest for delivering a high tumor dose safely in a single fraction. To support needle placement in the limited workspace inside the closed-bore MRI, a single-needle MR-compatible robot is currently under development at the University Medical Center Utrecht (UMCU). This robotic device taps the needle in a divergent way from a single rotation point into the prostate. With this setup, it is warranted to deliver the irradiation dose by successive insertions of the needle. Although robot-assisted needle placement is expected to be more accurate than manual template-guided insertion, needle positioning errors may occur and are likely to modify the pre-planned dose distribution.In this paper, we propose a dose plan adaptation strategy for HDR prostate brachytherapy with feedback on the needle position: a dose plan is made at the beginning of the interventional procedure and updated after each needle insertion in order to compensate for possible needle positioning errors. The introduced procedure can be used with the single needle MR-compatible robot developed at the UMCU. The proposed feedback strategy was tested by simulating complete HDR procedures with and without feedback on eight patients with different numbers of needle insertions (varying from 4 to 12). In of the cases tested, the number of clinically acceptable plans obtained at the end of the procedure was larger with feedback compared to the situation without feedback. Furthermore, the computation time of the feedback between each insertion was below 100 s which makes it eligible for intra-operative use.

Development and internal validation of a multivariable prediction model for biochemical failure after whole-gland salvage iodine-125 prostate brachytherapy for recurrent prostate cancer

BACKGROUND

Localized recurrent prostate cancer after primary radiotherapy can be curatively treated using salvage iodine-125 ((125)I) brachytherapy. Selection is hampered by a lack of predictive factors for cancer control. This study aims to develop and internally validate a prognostic model for biochemical failure (BF) after salvage (125)I brachytherapy.

METHODS AND MATERIALS

Whole-gland salvage (125)I brachytherapy patients were treated between 1993 and 2010 in two radiotherapy centers in the Netherlands. Multivariable Cox regression was performed to assess the predictive value of clinical parameters related to BF (Phoenix-definition [prostate-specific antigen [PSA]-nadir + 2.0 ng/mL]). Missing data were handled by multiple imputation. The model's discriminatory ability was assessed with Harrell's C-statistic. Internal validation was performed using bootstrap resampling (2000 data sets). Goodness-of-fit was evaluated with calibration plots. All analyses were performed using the recently published TRIPOD (Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis) statement.

RESULTS

After median followup of 74 months (range 5-138), 43 of a total 62 patients developed BF. In multivariable analysis, disease-free survival interval (DFSI) after primary therapy and pre-salvage prostate-specific antigen doubling time (PSADT) were predictors of BF: corrected hazard ratio (HR) 0.99 (95% confidence interval 0.97-0.999; p = 0.04) and 0.94 (95% confidence interval 0.89-0.99; p = 0.03), both for a 1-month increase (optimism-adjusted C-statistic 0.70). Calibration was accurate up to 36 months. Of patients with PSADT >30 months and DFSI >60 months, 36-month biochemical disease-free survival was >75%. Every 12-month increase in DFSI will allow 3-month decrease in PSADT while maintaining the same biochemical recurrence-free rates.

CONCLUSIONS

We have presented results from a cohort of patients undergoing salvage (125)I-brachytherapy. Our data show that better selection of patients is possible with the DFSI and PSADT.

An automated optimization tool for high-dose-rate (HDR) prostate brachytherapy with divergent needle pattern

Focal high-dose-rate (HDR) for prostate cancer has gained increasing interest as an alternative to whole gland therapy as it may contribute to the reduction of treatment related toxicity. For focal treatment, optimal needle guidance and placement is warranted. This can be achieved under MR guidance. However, MR-guided needle placement is currently not possible due to space restrictions in the closed MR bore. To overcome this problem, a MR-compatible, single-divergent needle-implant robotic device is under development at the University Medical Centre, Utrecht: placed between the legs of the patient inside the MR bore, this robot will tap the needle in a divergent pattern from a single rotation point into the tissue. This rotation point is just beneath the perineal skin to have access to the focal prostate tumor lesion. Currently, there is no treatment planning system commercially available which allows optimization of the dose distribution with such needle arrangement. The aim of this work is to develop an automatic inverse dose planning optimization tool for focal HDR prostate brachytherapy with needle insertions in a divergent configuration. A complete optimizer workflow is proposed which includes the determination of (1) the position of the center of rotation, (2) the needle angulations and (3) the dwell times. Unlike most currently used optimizers, no prior selection or adjustment of input parameters such as minimum or maximum dose or weight coefficients for treatment region and organs at risk is required. To test this optimizer, a planning study was performed on ten patients (treatment volumes ranged from 8.5 cm(3)to 23.3 cm(3)) by using 2-14 needle insertions. The total computation time of the optimizer workflow was below 20 min and a clinically acceptable plan was reached on average using only four needle insertions.

Finite element modelling of contracting skeletal muscle

To describe the mechanical behaviour of biological tissues and transport processes in biological tissues, conservation laws such as conservation of mass, momentum and energy play a central role. Mathematically these are cast into the form of partial differential equations. Because of nonlinear material behaviour, inhomogeneous properties and usually a complex geometry, it is impossible to find closed-form analytical solutions for these sets of equations. The objective of the finite element method is to find approximate solutions for these problems. The concepts of the finite element method are explained on a finite element continuum model of skeletal muscle. In this case, the momentum equations have to be solved with an extra constraint, because the material behaves as nearly incompressible. The material behaviour consists of a highly nonlinear passive part and an active part. The latter is described with a two-state Huxley model. This means that an extra nonlinear partial differential equation has to be solved. The problems and solutions involved with this procedure are explained. The model is used to describe the mechanical behaviour of a tibialis anterior of a rat. The results have been compared with experimentally determined strains at the surface of the muscle. Qualitatively there is good agreement between measured and calculated strains, but the measured strains were higher.

Spatial and temporal heterogeneity of superficial muscle strain during in situ fixed-end contractions

Numerical models of contracting muscle offer a powerful tool to study local mechanical load. For validation of these models, the spatial and temporal distribution of strain was quantified in fixed-end contracting rat tibialis anterior muscle in situ at optimal muscle length (L(o)) and at 120 degrees plantar flexion as well as at 125 and 33Hz stimulation frequency. We studied the hypothesis that after termination of stimulation in situ muscle segments near the motor endplates elongate while segments away from the endplates shorten. We show that both spatial and temporal inhomogeneities in muscle deformation occurred during contraction. Muscle plateau shortening strain equalled 4.1%. Maximal plateau shortening of a muscle segment was much larger (9.6%) and occurred distally (at 0.26 of the scaled length of the muscle). Manipulating torque levels by decreasing the stimulation frequency at the same muscle length induced a decrease in torque ( approximately 20%) with a smaller effect on the level and no effect on the pattern of muscle deformation. During relaxation, distal segments actively shortened at the expense of proximal muscle segments, which elongated. The segments undergoing lengthening were nearer to motor endplates than segments undergoing shortening. In conclusion, the present study provides experimental data on magnitude of contraction-induced deformation needed for validation of numerical models. Local muscle deformation is heterogeneous both temporally and spatially and may be related to proximity to the motor endplates.