Evaluation of scatter-to-primary ratio in radiological conditions

Quantifying the effect of X-ray scattering for data generation in real-time defect detection

BACKGROUND

X-ray imaging is widely used for the non-destructive detection of defects in industrial products on a conveyor belt. In-line detection requires highly accurate, robust, and fast algorithms. Deep Convolutional Neural Networks (DCNNs) satisfy these requirements when a large amount of labeled data is available. To overcome the challenge of collecting these data, different methods of X-ray image generation are considered.

OBJECTIVE

Depending on the desired degree of similarity to real data, different physical effects should either be simulated or can be ignored. X-ray scattering is known to be computationally expensive to simulate, and this effect can greatly affect the accuracy of a generated X-ray image. We aim to quantitatively evaluate the effect of scattering on defect detection.

METHODS

Monte-Carlo simulation is used to generate X-ray scattering distribution. DCNNs are trained on the data with and without scattering and applied to the same test datasets. Probability of Detection (POD) curves are computed to compare their performance, characterized by the size of the smallest detectable defect.

RESULTS

We apply the methodology to a model problem of defect detection in cylinders. When trained on data without scattering, DCNNs reliably detect defects larger than 1.3 mm, and using data with scattering improves performance by less than 5%. If the analysis is performed on the cases with large scattering-to-primary ratio (1 < SPR < 5), the difference in performance could reach 15% (approx. 0.4 mm).

CONCLUSION

Excluding the scattering signal from the training data has the largest effect on the smallest detectable defects, and the difference decreases for larger defects. The scattering-to-primary ratio has a significant effect on detection performance and the required accuracy of data generation.

Boron neutron capture therapy (BNCT) for liver metastasis in an experimental model: dose–response at five-week follow-up based on retrospective dose assessment in individual rats

Boron neutron capture therapy (BNCT) was proposed for untreatable colorectal liver metastases. Employing an experimental model of liver metastases in rats, we recently demonstrated that BNCT mediated by boronophenylalanine (BPA-BNCT) at 13 Gy prescribed to tumor is therapeutically useful at 3-week follow-up. The aim of the present study was to evaluate dose–response at 5-week follow-up, based on retrospective dose assessment in individual rats. BDIX rats were inoculated with syngeneic colon cancer cells DHD/K12/TRb. Tumor-bearing animals were divided into three groups: BPA-BNCT (n = 19), Beam only (n = 8) and Sham (n = 7) (matched manipulation, no treatment). For each rat, neutron flux was measured in situ and boron content was measured in a pre-irradiation blood sample for retrospective individual dose assessment. For statistical analysis (ANOVA), individual data for the BPA-BNCT group were pooled according to absorbed tumor dose, BPA-BNCT I: 4.5–8.9 Gy and BPA-BNCT II: 9.2–16 Gy. At 5 weeks post-irradiation, the tumor surface area post-treatment/pre-treatment ratio was 12.2 ± 6.6 for Sham, 7.8 ± 4.1 for Beam only, 4.4 ± 5.6 for BPA-BNCT I and 0.45 ± 0.20 for BPA-BNCT II; tumor nodule weight was 750 ± 480 mg for Sham, 960 ± 620 mg for Beam only, 380 ± 720 mg for BPA-BNCT I and 7.3 ± 5.9 mg for BPA-BNCT II. The BPA-BNCT II group exhibited statistically significant tumor control with no contributory liver toxicity. Potential threshold doses for tumor response and significant tumor control were established at 6.1 and 9.2 Gy, respectively.

Boron delivery with liposomes for boron neutron capture therapy (BNCT): biodistribution studies in an experimental model of oral cancer demonstrating therapeutic potential

Boron neutron capture therapy (BNCT) combines selective accumulation of (10)B carriers in tumor tissue with subsequent neutron irradiation. We previously demonstrated the therapeutic efficacy of BNCT in the hamster cheek pouch oral cancer model. Optimization of BNCT depends largely on improving boron targeting to tumor cells. Seeking to maximize the potential of BNCT for the treatment for head and neck cancer, the aim of the present study was to perform boron biodistribution studies in the oral cancer model employing two different liposome formulations that were previously tested for a different pathology, i.e., in experimental mammary carcinoma in BALB/c mice: (1) MAC: liposomes incorporating K[nido-7-CH(3)(CH(2))(15)-7,8-C(2)B(9)H(11)] in the bilayer membrane and encapsulating a hypertonic buffer, administered intravenously at 6 mg B per kg body weight, and (2) MAC-TAC: liposomes incorporating K[nido-7-CH(3)(CH(2))(15)-7,8-C(2)B(9)H(11)] in the bilayer membrane and encapsulating a concentrated aqueous solution of the hydrophilic species Na(3) [ae-B(20)H(17)NH(3)], administered intravenously at 18 mg B per kg body weight. Samples of tumor, precancerous and normal pouch tissue, spleen, liver, kidney, and blood were taken at different times post-administration and processed to measure boron content by inductively coupled plasma mass spectrometry. No ostensible clinical toxic effects were observed with the selected formulations. Both MAC and MAC-TAC delivered boron selectively to tumor tissue. Absolute tumor values for MAC-TAC peaked to 66.6 ± 16.1 ppm at 48 h and to 43.9 ± 17.6 ppm at 54 h with very favorable ratios of tumor boron relative to precancerous and normal tissue, making these protocols particularly worthy of radiobiological assessment. Boron concentration values obtained would result in therapeutic BNCT doses in tumor without exceeding radiotolerance in precancerous/normal tissue at the thermal neutron facility at RA-3.