Improved outcome models with denoising diffusion.

PURPOSE: Radiotherapy outcome modelling often suffers from class imbalance in the modelled endpoints. One of the main options to address this issue is by introducing new synthetically generated datapoints, using generative models, such as Denoising Diffusion Probabilistic Models (DDPM). In this study, we implemented DDPM to improve performance of a tumor local control model, trained on imbalanced dataset, and compare this approach with other common techniques. METHODS: A dataset of 535 NSCLC patients treated with SBRT (50 Gy/5 fractions) was used to train a deep learning outcome model for tumor local control prediction. The dataset included complete treatment planning data (planning CT images, 3D planning dose distribution and patient demographics) with sparsely distributed endpoints (6-7 % experiencing local failure). Consequently, we trained a novel conditional 3D DDPM model to generate synthetic treatment planning data. Synthetically generated treatment planning datapoints were used to supplement the real training dataset and the improvement in the model's performance was studied. Obtained results were also compared to other common techniques for class imbalanced training, such as Oversampling, Undersampling, Augmentation, Class Weights, SMOTE and ADASYN. RESULTS: Synthetic DDPM-generated data were visually trustworthy, with Fréchet inception distance (FID) below 50. Extending the training dataset with the synthetic data improved the model's performance by more than 10%, while other techniques exhibited only about 4% improvement. CONCLUSIONS: DDPM introduces a novel approach to class-imbalanced outcome modelling problems. The model generates realistic synthetic radiotherapy planning data, with a strong potential to increase performance and robustness of outcome models.

High-resolution MRI synthesis using a data-driven framework with denoising diffusion probabilistic modeling.

Objective. High-resolution magnetic resonance imaging (MRI) can enhance lesion diagnosis, prognosis, and delineation. However, gradient power and hardware limitations prohibit recording thin slices or sub-1 mm resolution. Furthermore, long scan time is not clinically acceptable. Conventional high-resolution images generated using statistical or analytical methods include the limitation of capturing complex, high-dimensional image data with intricate patterns and structures. This study aims to harness cutting-edge diffusion probabilistic deep learning techniques to create a framework for generating high-resolution MRI from low-resolution counterparts, improving the uncertainty of denoising diffusion probabilistic models (DDPM).Approach. DDPM includes two processes. The forward process employs a Markov chain to systematically introduce Gaussian noise to low-resolution MRI images. In the reverse process, a U-Net model is trained to denoise the forward process images and produce high-resolution images conditioned on the features of their low-resolution counterparts. The proposed framework was demonstrated using T2-weighted MRI images from institutional prostate patients and brain patients collected in the Brain Tumor Segmentation Challenge 2020 (BraTS2020).Main results. For the prostate dataset, the bicubic interpolation model (Bicubic), conditional generative-adversarial network (CGAN), and our proposed DDPM framework improved the noise quality measure from low-resolution images by 4.4%, 5.7%, and 12.8%, respectively. Our method enhanced the signal-to-noise ratios by 11.7%, surpassing Bicubic (9.8%) and CGAN (8.1%). In the BraTS2020 dataset, the proposed framework and Bicubic enhanced peak signal-to-noise ratio from resolution-degraded images by 9.1% and 5.8%. The multi-scale structural similarity indexes were 0.970 ± 0.019, 0.968 ± 0.022, and 0.967 ± 0.023 for the proposed method, CGAN, and Bicubic, respectively.Significance. This study explores a deep learning-based diffusion probabilistic framework for improving MR image resolution. Such a framework can be used to improve clinical workflow by obtaining high-resolution images without penalty of the long scan time. Future investigation will likely focus on prospectively testing the efficacy of this framework with different clinical indications.

CBCT-Based synthetic CT image generation using conditional denoising diffusion probabilistic model.

BACKGROUND: Daily or weekly cone-beam computed tomography (CBCT) scans are commonly used for accurate patient positioning during the image-guided radiotherapy (IGRT) process, making it an ideal option for adaptive radiotherapy (ART) replanning. However, the presence of severe artifacts and inaccurate Hounsfield unit (HU) values prevent its use for quantitative applications such as organ segmentation and dose calculation. To enable the clinical practice of online ART, it is crucial to obtain CBCT scans with a quality comparable to that of a CT scan. PURPOSE: This work aims to develop a conditional diffusion model to perform image translation from the CBCT to the CT distribution for the image quality improvement of CBCT. METHODS: The proposed method is a conditional denoising diffusion probabilistic model (DDPM) that utilizes a time-embedded U-net architecture with residual and attention blocks to gradually transform the white Gaussian noise sample to the target CT distribution conditioned on the CBCT. The model was trained on deformed planning CT (dpCT) and CBCT image pairs, and its feasibility was verified in brain patient study and head-and-neck (H&N) patient study. The performance of the proposed algorithm was evaluated using mean absolute error (MAE), peak signal-to-noise ratio (PSNR) and normalized cross-correlation (NCC) metrics on generated synthetic CT (sCT) samples. The proposed method was also compared to four other diffusion model-based sCT generation methods. RESULTS: In the brain patient study, the MAE, PSNR, and NCC of the generated sCT were 25.99 HU, 30.49 dB, and 0.99, respectively, compared to 40.63 HU, 27.87 dB, and 0.98 of the CBCT images. In the H&N patient study, the metrics were 32.56 HU, 27.65 dB, 0.98 and 38.99 HU, 27.00, 0.98 for sCT and CBCT, respectively. Compared to the other four diffusion models and one Cycle generative adversarial network (Cycle GAN), the proposed method showed superior results in both visual quality and quantitative analysis. CONCLUSIONS: The proposed conditional DDPM method can generate sCT from CBCT with accurate HU numbers and reduced artifacts, enabling accurate CBCT-based organ segmentation and dose calculation for online ART.

Structural changes in subdomain 2 of G-actin observed by fluorescence spectroscopy.

The influence of DNase I binding to Ca-ATP-G-actin and of Ca2+/Mg2+ and ATP/ADP exchange on the conformation of G-actin were investigated by measuring the fluorescence of dansyl cadaverine (DC) conjugated to Gln41 in subdomain 2 of the protein. Fluorescence resonance energy transfer (FRET) between this probe and N-[4-(dimethylamino)-3,5-dinitrophenyl]maleimide (DDPM) attached to Cys374 in subdomain 1 was also measured. Contrary to an earlier report [dos Remedios, Kiessling and Hambly (1994) in Synchrotron Radiation in the Biosciences (Chance, B., Deisenhofer, J., Ebashi, S., Goodhead, D. T., Helliwell, J. R., Huxley, H. E., Iizuka, T., Kirz, J., Mitsui, T., Rubenstein, E. et al., eds.), pp. 418-425, Oxford University Press, Oxford], the distance between these probes did not change significantly when DNase I was bound to actin. A small but reproducible increase in the quantum yield and a blue shift of the DC fluorescence maximum were observed when bound Ca2+ was replaced by Mg2+. A large increase (about 70%) in the quantum yield and an approx. 12 nm blue shift of the emission spectrum occurred when ATP in Mg-G-actin was replaced by ADP. These changes were not accompanied by any significant change in the FRET distance between the dansyl donor and DDPM acceptor probes. A substantial change in the fluorescence of DC-actin was observed after proteolytic removal of the last three residues of actin, in accordance with earlier evidence suggesting that there is a conformational coupling between subdomain 2 and the C-terminal segment in subdomain 1 of actin. The results are discussed in relation to recently published data obtained with another fluorescent probe and to earlier observations based on limited cleavage using proteolytic enzymes.

Energy-transfer measurements of the Cys35-Cys84 distance in bovine cardiac troponin C.

Bovine cardiac troponin C (cTnC) has cysteine residues located in the non-functional Ca(2+)-binding loop I (Cys-35) and at the N-terminal end of the central helix (Cys-84) near site II, the regulatory Ca(2+)-binding site. Recently, we reported that the excimer fluorescence resulting from the dimerization of adjacent pyrene groups attached to the two Cys residues is reduced by Ca2+ binding to site II (Liou, Y.-M. and Fuchs, F. (1992) Biophys. J. 61, 892-901). This result would suggest that Ca2+ binding causes a separation of the two Cys residues, a conclusion at variance with predictions from molecular modeling studies (Herzberg, O., Moult, J. and James, M.N.G. (1986) J. Biol. Chem. 261, 2638-2644). Alternatively, the reduction in excimer fluorescence could be accounted for by an immobilization of the pyrene attached to Cys-84 by a Ca(2+)-induced hydrophobic pocket. To arrive at a more definitive interpretation of these experiments, we carried out steady-state fluorescence resonance energy-transfer measurements of the Cys35-Cys84 distance. We used three different donor-acceptor pairs: 2-(4'-(iodoacetamido)anilino) naphthalene-6-sulfonic acid (IAANS) and 4-dimethylamino-phenylazophenyl-4-maleimide (DABMI), IAANS and N-(4-(dimethyl-amino)-3,5-dinitrophenyl) maleimide (DDPM), and 5-((((2-iodoacetyl)amino)ethyl)amino)naphthalene-1-sulfonic acid (IAEDANS) and DDPM. At pCa 8.0, the distances were 23.8, 21.0, and 22.0 A with the donor-acceptor pairs, IAANS-DABMI, IAANS-DDPM and IAEDAN-DDPM, respectively. At pCa 4.0, the distances were 25.8, 24.1 and 21.2 A. The distances at pCa 8 and pCa 4.0 were not significantly altered when labeled cTnC was complexed with cardiac troponin I (cTnI). Thus, Ca2+ has little, if any, effect on the Cys35-Cys84 distance. These results are consistent with a model in which Ca2+ binding induces a separation of helices B and C from helix D, without any relative movement of the two N-terminal Ca(2+)-binding domains.

Location of the essential cysteine residue of jack bean urease.

Jack bean urease is inactivated by the modification of about one cysteine residue per subunit with N-ethylmaleimide (NEM) or other thiol titrant. The location of this cysteine residue was identified. After blocking the unessential thiol groups with NEM, the essential cysteine was labeled with N-(4-dimethylamino-3,5-dinitrophenyl)maleimide (DDPM). The DDPM-labeled protein was cleaved with cyanogen bromide and a DDPM-fragment was purified by gel filtration and ion exchange chromatography. Amino-terminal amino acid sequence of the DDPM-labeled fragment corresponded to that of a cyanogen bromide fragment of urease, VCHHLDREIPEDLAFAHSRIRKKTIAAEDVLNDIGAISIISSDSQAM. The second residue, Cys-592 in native urease, was labeled with DDPM. Fluoride ion, which competitively inhibits urease activity, protected urease from inactivation by NEM. The dissociation constant of fluoride ion obtained from the rate of inactivation with NEM was essentially identical to both the kinetically and spectrophotometrically determined dissociation constants. These suggest that the essential cysteine is at or near the active site.

Fluorescence resonance energy transfer between the nucleotide binding site and Cys-10 in G-actin and F-actin.

Intramonomer fluorescence resonance energy transfer between the donor epsilon-ATP bound to the nucleotide site and the acceptor N-(4-dimethylamino-3,5-dinitrophenyl)maleimide (DDPM) or 4-dimethylaminophenyl-azophenyl-4'-maleimide bound to Cys-10 in G-actin was measured. The donor-acceptor distance was calculated to be about 40 A. The intermonomer energy transfer in F-actin occurring between epsilon-ADP and DABMI was also measured. The radial coordinate of Cys-10 was calculated to be 25 A based on the helical symmetry of F-actin and the recently calculated radial coordinate of the nucleotide binding site in F-actin i.e. 25 A (Miki, M., Hambly, B. and dos Remedios, C.G. (1986) Biochim. Biophys. Acta 871, 137-141). (The assumption has been made in calculating these distances that the energy donor and acceptor rotate rapidly relative to the fluorescence lifetime.) Corresponding distances separating the donor nucleotide in one monomer from acceptors on Cys-10 in the first and second nearest neighbours in F-actin are 39-40 A and 41-43 A.

Modification by antioxidants and p,p'-diaminodiphenylmethane of 7,12-dimethylbenz[a]anthracene-induced carcinogenesis of the mammary gland and ear duct in CD rats.

The effects of antioxidants on mammary gland carcinogenesis pretreated with 7,12-dimethylbenz[a]anthracene (DMBA) in female Sprague-Dawley rats were examined. The antioxidants used were butylated hydroxyanisole (BHA), butylated hydroxytoluene (BHT), sodium L-ascorbate, alpha-tocopherol, ethoxyquin and p,p'-diaminodiphenylmethane (DDPM), which is an inhibitor of carcinogenesis in the liver, kidney and urinary bladder. Female Sprague-Dawley rats of 50 days old were treated with 2.5 mg/100 g body weight of DMBA, and from 1 week later were given diet supplemented with 1% BHA, 0.7% BHT, 5% sodium L-ascorbate, 1.5% alpha-tocopherol, 0.5% ethoxyquin or 0.1% DDPM for 33 weeks and then killed. The incidences of mammary tumors, carcinomas and fibroadenomas in DMBA-treated animals were reduced by diet containing BHA or ethoxyquin. Diet containing BHT or DDPM inhibited the induction of only fibroadenomas. The incidence of ear duct tumors in DMBA-treated animals was reduced by diet containing BHT, alpha-tocopherol or ethoxyquin.

Qualitative and quantitative analysis of some synthetic, chemically acting laxatives in urine by gas chromatography-mass spectrometry.

A method for the qualitative and quantitative simultaneous analysis of dioxyanthraquinone, desacetyl-Bisacodyl, phenolphthalein and Oxyphenisatin in human urine using gas chromatography-mass spectrometry (GC-MS) has been developed. The compounds were extracted from urine at pH 7.5 with diethyl ether using Extrelut extraction columns, followed by evaporation and trimethylsilylation. The method used electron beam ionization GC-MS employing a computer-controlled multiple-ion detector (mass fragmentography). The recovery from urine for the various compounds was between 80% and 100%. The detection limit for these compounds was in the range 0.01--0.05 micrograms/ml of urine. The method proved to be suitable for measuring urine concentrations for at least four days after administration of a single oral low therapeutic dose of the laxatives to sixteen healthy volunteers.