Automated Protocoling for MRI Exams-Challenges and Solutions.

Automated protocoling for MRI examinations is an amendable target for workflow automation with artificial intelligence. However, there are still challenges to overcome for a successful and robust approach. These challenges are outlined and analyzed in this work. Through a literature review, we analyzed limitations of currently published approaches for automated protocoling. Then, we assessed these limitations quantitatively based on data from a private radiology practice. For this, we assessed the information content provided by the clinical indication by computing the overlap coefficients for the sets of ICD-10-coded admitting diagnoses of different MRI protocols. Additionally, we assessed the heterogeneity of protocol trees from three different MRI scanners based on the overlap coefficient, on MRI protocol and sequence level. Additionally, we applied sequence name standardization to demonstrate its effect on the heterogeneity assessment, i.e., the overlap coefficient, of different protocol trees. The overlap coefficient for the set of ICD-10-coded admitting diagnoses for different protocols ranges from 0.14 to 0.56 for brain/head MRI exams and 0.04 to 0.57 for spine exams. The overlap coefficient across the set of sequences used at two different scanners increases when applying sequence name standardization (from 0.81/0.86 to 0.93). Automated protocoling for MRI examinations has the potential to reduce the workload for radiologists. However, an automated protocoling approach cannot be solely based on admitting diagnosis as it does not provide sufficient information. Moreover, sequence name standardization increases the overlap coefficient across the set of sequences used at different scanners and therefore facilitates transfer learning.

Detection of Ablation Boundaries Using Different MR Sequences in a Swine Liver Model.

PURPOSE: To determine the magnetic resonance (MR) sequences best suited for the assessment of ablation zones after radiofrequency ablation (RFA). METHODS: Three percutaneous MR-guided RFA of the liver were performed on three swine. Four pre-contrast and two hepatobiliary post-contrast sequences were obtained after ablation. Tissue samples were extracted and stained for nicotinamide adenine dinucleotide diaphorase hydride (NADH) and with hematoxylin and eosin. Post-ablation MR images and NADH slides were segmented to determine the total ablation zone, their Dice similarity coefficient (DSC), and the contrast-to-noise ratio (CNR) of the visible ablation boundary to normal liver tissue. RESULTS: Two distinct layers were combined to determine the ablation zone: an inner layer of coagulation necrosis and an outer layer defined as the peripheral transition zone. Corresponding zones could be found in the MR images as well. Compared to histology, the total area of the MR ablation zone was significantly smaller on the pre-contrast T1 images (p < 0.01) and significantly larger with T2 turbo spin-echo (p = 0.025). No significant difference in size of the ablation zone depiction could be found between histology, post-contrast T1 volumetric interpolated breath-hold examination (VIBE), and post-contrast T1 3D Turboflash (TFL) as well as T2 SPACE images. All sequences but the pre-contrast T1 VIBE sequence showed a DSC above 80% and a high CNR. CONCLUSIONS: Post-contrast T1 3DTFL performs best when assessing ablation zones after RFA. Since the sequence requires a long acquisition time, T1 VIBE post-contrast offers the best compromise between acquisition time and estimation accuracy.

Rule-Based Models for Risk Estimation and Analysis of In-hospital Mortality in Emergency and Critical Care.

We propose a novel method that uses associative classification and odds ratios to predict in-hospital mortality in emergency and critical care. Manual mortality risk scores have previously been used to assess the care needed for each patient and their need for palliative measures. Automated approaches allow providers to get a quick and objective estimation based on electronic health records. We use association rule mining to find relevant patterns in the dataset. The odds ratio is used instead of classical association rule mining metrics as a quality measure to analyze association instead of frequency. The resulting measures are used to estimate the in-hospital mortality risk. We compare two prediction models: one minimal model with socio-demographic factors that are available at the time of admission and can be provided by the patients themselves, namely gender, ethnicity, type of insurance, language, and marital status, and a full model that additionally includes clinical information like diagnoses, medication, and procedures. The method was tested and validated on MIMIC-IV, a publicly available clinical dataset. The minimal prediction model achieved an area under the receiver operating characteristic curve value of 0.69, while the full prediction model achieved a value of 0.98. The models serve different purposes. The minimal model can be used as a first risk assessment based on patient-reported information. The full model expands on this and provides an updated risk assessment each time a new variable occurs in the clinical case. In addition, the rules in the models allow us to analyze the dataset based on data-backed rules. We provide several examples of interesting rules, including rules that hint at errors in the underlying data, rules that correspond to existing epidemiological research, and rules that were previously unknown and can serve as starting points for future studies.

Predicting Anxiety in Routine Palliative Care Using Bayesian-Inspired Association Rule Mining.

We propose a novel knowledge extraction method based on Bayesian-inspired association rule mining to classify anxiety in heterogeneous, routinely collected data from 9,924 palliative patients. The method extracts association rules mined using lift and local support as selection criteria. The extracted rules are used to assess the maximum evidence supporting and rejecting anxiety for each patient in the test set. We evaluated the predictive accuracy by calculating the area under the receiver operating characteristic curve (AUC). The evaluation produced an AUC of 0.89 and a set of 55 atomic rules with one item in the premise and the conclusion, respectively. The selected rules include variables like pain, nausea, and various medications. Our method outperforms the previous state of the art (AUC = 0.72). We analyzed the relevance and novelty of the mined rules. Palliative experts were asked about the correlation between variables in the data set and anxiety. By comparing expert answers with the retrieved rules, we grouped rules into expected and unexpected ones and found several rules for which experts' opinions and the data-backed rules differ, most notably with the patients' sex. The proposed method offers a novel way to predict anxiety in palliative settings using routinely collected data with an explainable and effective model based on Bayesian-inspired association rule mining. The extracted rules give further insight into potential knowledge gaps in the palliative care field.

Enhanced Magnetic Resonance Image Synthesis with Contrast-Aware Generative Adversarial Networks.

A magnetic resonance imaging (MRI) exam typically consists of the acquisition of multiple MR pulse sequences, which are required for a reliable diagnosis. With the rise of generative deep learning models, approaches for the synthesis of MR images are developed to either synthesize additional MR contrasts, generate synthetic data, or augment existing data for AI training. While current generative approaches allow only the synthesis of specific sets of MR contrasts, we developed a method to generate synthetic MR images with adjustable image contrast. Therefore, we trained a generative adversarial network (GAN) with a separate auxiliary classifier (AC) network to generate synthetic MR knee images conditioned on various acquisition parameters (repetition time, echo time, and image orientation). The AC determined the repetition time with a mean absolute error (MAE) of 239.6 ms, the echo time with an MAE of 1.6 ms, and the image orientation with an accuracy of 100%. Therefore, it can properly condition the generator network during training. Moreover, in a visual Turing test, two experts mislabeled 40.5% of real and synthetic MR images, demonstrating that the image quality of the generated synthetic and real MR images is comparable. This work can support radiologists and technologists during the parameterization of MR sequences by previewing the yielded MR contrast, can serve as a valuable tool for radiology training, and can be used for customized data generation to support AI training.

Automating Time-Consuming and Error-Prone Manual Nursing Management Documentation Processes.

A German regulation requires nursing managers to document patient-nurse ratios. They have to combine heterogeneous hospital data from different sources. Missing documentation or ratios that are too high lead to sanctions. Automated approaches are needed to accelerate the time-consuming and error-prone documentation process. A documentation and visualization system was implemented. The system allows nursing managers to quickly and automatically create the documentation required by the regulation. Interactive visualization dashboards assist with the analysis of patient and staff numbers. The developed method was effectively used in nursing management tasks. No changes to the information technology infrastructure were needed. The new process is around 35 hours per month faster and less error-prone. The documentation functionality automatically reads the required information and correctly calculates the documentation. The visualization functionality allows nursing managers to assess the current patient-nurse ratios before the documentation is submitted. The method scales to multiple wards and locations. It calculates the sanctions to expect and is easily updatable. The proposed method is expected to decrease nursing administration workloads and facilitate the analysis of nursing management data in a cost-effective way.

MR-contrast-aware image-to-image translations with generative adversarial networks.

PURPOSE: A magnetic resonance imaging (MRI) exam typically consists of several sequences that yield different image contrasts. Each sequence is parameterized through multiple acquisition parameters that influence image contrast, signal-to-noise ratio, acquisition time, and/or resolution. Depending on the clinical indication, different contrasts are required by the radiologist to make a diagnosis. As MR sequence acquisition is time consuming and acquired images may be corrupted due to motion, a method to synthesize MR images with adjustable contrast properties is required. METHODS: Therefore, we trained an image-to-image generative adversarial network conditioned on the MR acquisition parameters repetition time and echo time. Our approach is motivated by style transfer networks, whereas the "style" for an image is explicitly given in our case, as it is determined by the MR acquisition parameters our network is conditioned on. RESULTS: This enables us to synthesize MR images with adjustable image contrast. We evaluated our approach on the fastMRI dataset, a large set of publicly available MR knee images, and show that our method outperforms a benchmark pix2pix approach in the translation of non-fat-saturated MR images to fat-saturated images. Our approach yields a peak signal-to-noise ratio and structural similarity of 24.48 and 0.66, surpassing the pix2pix benchmark model significantly. CONCLUSION: Our model is the first that enables fine-tuned contrast synthesis, which can be used to synthesize missing MR-contrasts or as a data augmentation technique for AI training in MRI. It can also be used as basis for other image-to-image translation tasks within medical imaging, e.g., to enhance intermodality translation (MRI → CT) or 7 T image synthesis from 3 T MR images.

Machine Learning-Based HIV Risk Estimation Using Incidence Rate Ratios.

HIV/AIDS is an ongoing global pandemic, with an estimated 39 million infected worldwide. Early detection is anticipated to help improve outcomes and prevent further infections. Point-of-care diagnostics make HIV/AIDS diagnoses available both earlier and to a broader population. Wide-spread and automated HIV risk estimation can offer objective guidance. This supports providers in making an informed decision when considering patients with high HIV risk for HIV testing or pre-exposure prophylaxis (PrEP). We propose a novel machine learning method that allows providers to use the data from a patient's previous stays at the clinic to estimate their HIV risk. All features available in the clinical data are considered, making the set of features objective and independent of expert opinions. The proposed method builds on association rules that are derived from the data. The incidence rate ratio (IRR) is determined for each rule. Given a new patient, the mean IRR of all applicable rules is used to estimate their HIV risk. The method was tested and validated on the publicly available clinical database MIMIC-IV, which consists of around 525,000 hospital stays that included a stay at the intensive care unit or emergency department. We evaluated the method using the area under the receiver operating characteristic curve (AUC). The best performance with an AUC of 0.88 was achieved with a model consisting of 53 rules. A threshold value of 0.66 leads to a sensitivity of 98% and a specificity of 53%. The rules were grouped into drug abuse, psychological illnesses (e.g., PTSD), previously known associations (e.g., pulmonary diseases), and new associations (e.g., certain diagnostic procedures). In conclusion, we propose a novel HIV risk estimation method that builds on existing clinical data. It incorporates a wide range of features, leading to a model that is independent of expert opinions. It supports providers in making informed decisions in the point-of-care diagnostics process by estimating a patient's HIV risk.

Automated Billing Code Retrieval from MRI Scanner Log Data.

Although the level of digitalization and automation steadily increases in radiology, billing coding for magnetic resonance imaging (MRI) exams in the radiology department is still based on manual input from the technologist. After the exam completion, the technologist enters the corresponding exam codes that are associated with billing codes in the radiology information system. Moreover, additional billing codes are added or removed, depending on the performed procedure. This workflow is time-consuming and we showed that billing codes reported by the technologists contain errors. The coding workflow can benefit from an automated system, and thus a prediction model for automated assignment of billing codes for MRI exams based on MRI log data is developed in this work. To the best of our knowledge, it is the first attempt to focus on the prediction of billing codes from modality log data. MRI log data provide a variety of information, including the set of executed MR sequences, MR scanner table movements, and given a contrast medium. MR sequence names are standardized using a heuristic approach and incorporated into the features for the prediction. The prediction model is trained on 9754 MRI exams and tested on 1 month of log data (423 MRI exams) from two MRI scanners of the radiology site for the Swiss medical tariffication system Tarmed. The developed model, an ensemble of classifier chains with multilayer perceptron as a base classifier, predicts medical billing codes for MRI exams with a micro-averaged F1-score of 97.8% (recall 98.1%, precision 97.5%). Manual coding reaches a micro-averaged F1-score of 98.1% (recall 97.4%, precision 98.8%). Thus, the performance of automated coding is close to human performance. Integrated into the clinical environment, this work has the potential to free the technologist from a non-value adding an administrative task, therefore enhance the MRI workflow, and prevent coding errors.

Fast 3-T MR-guided transrectal prostate biopsy using an in-room tablet device for needle guide alignment: a feasibility study.

OBJECTIVES: To assess the feasibility of adding a tablet device inside the scanner room to assist needle-guide alignment during magnetic resonance (MR)-guided transrectal prostate biopsy. METHODS: Twenty patients with one cancer-suspicious region (CSR) with PI-RADS score ≥ 4 on diagnostic multiparametric MRI were prospectively enrolled. Two orthogonal scan planes of an MR fluoroscopy sequence (~3 images/s) were aligned to the CSR and needle-guide pivoting point. Targeting was achieved by manipulating the needle-guide under MR fluoroscopy feedback on the in-room tablet device. Technical feasibility and targeting success were assessed. Complications and biopsy procedure times were also recorded. RESULTS: Needle-guide alignment with the in-room tablet device was technically successful in all patients and allowed sampling after a single alignment step in 19/20 (95%) CSRs (median size 14 mm, range: 4-45). Biopsy cores contained cancer in 18/20 patients. There were no per-procedural or post-biopsy complications. Using the tablet device, the mean time to first biopsy was 5.8 ± 1.0 min and the mean total procedure time was 23.7 ± 4.1 min. CONCLUSIONS: Use of an in-room tablet device to assist needle-guide alignment was feasible and safe during MR-guided transrectal prostate biopsy. Initial experience indicates potential for procedure time reduction. KEY POINTS: • Performing MR-guided prostate biopsy using an in-room tablet device is feasible. • CSRs could be sampled after a single alignment step in 19/20 patients. • The mean procedure time for biopsy with the tablet device was 23.7 min.

Diffusion-weighted imaging during MR-guided radiofrequency ablation of hepatic malignancies: analysis of immediate pre- and post-ablative diffusion characteristics.

BACKGROUND: Previous studies have shown a benefit of magnetic resonance (MR)-diffusion-weighted imaging (DWI) for follow-up after liver radiofrequency (RF) ablation. However, no data are available concerning acute changes of DWI characteristics immediately after RF ablation. PURPOSE: To analyze and compare the MR-diffusion characteristics of pre-interventional hepatic malignancies and the ablation zone during successful MR-guided RF ablation. MATERIAL AND METHODS: This retrospective study was conducted in accordance with the guidelines of the local institutional review board. Forty-seven patients with 29 HCC (24 patients) and 30 hepatic metastases (23 patients) underwent MR-guided radiofrequency ablation including DWI before and immediately after ablation (b = 0, 400, 800 s/mm(2)). Two reviewers (A and B) analyzed DWI with focus on detectability of the tumor before ablation and characteristics of the coagulative area after treatment. Mean apparent diffusion coefficient (ADC) was compared between liver, untreated tumor, and hyperintense areas in post-ablative DWI (b = 800 s/mm(2)) with the paired Student's t-test. RESULTS: Pre-ablative: the reviewers classified 19/29 (A) and 23/29 (B) HCC and 25/30 (A and B) metastases as detectable in DWI. Post-ablative: a hyperintense rim surrounding the ablation zone was observed in 28/29 treated HCC and 30/30 treated metastases (A and B). A homogenous hypointense central ablation zone was found in 18/29 (A) and 20/29 (B) treated HCC and 17/30 (A & B) treated metastases in DWI. ADC of the rim was significantly lower than ADC of the liver (P < 0.001). CONCLUSION: DWI enables visualization of the target tumor in MR-guided liver radiofrequency ablation in most cases. A common post-ablative DWI finding is a hyperintense rim with decreased ADC surrounding the ablation zone.

Liver lesion conspicuity during real-time MR-guided radiofrequency applicator placement using spoiled gradient echo and balanced steady-state free precession imaging.

PURPOSE: To retrospectively evaluate the conspicuity of liver lesions in a fluoroscopic spoiled gradient echo (GRE) and a balanced steady-state free precession (SSFP) magnetic resonance imaging (MRI) sequence. MATERIALS AND METHODS: In all, 103 patients with hepatocellular carcinomas (HCC) (41) or liver metastases (67) were treated using MR-guided radiofrequency ablation in a wide-bore 1.5 T scanner. A multislice real-time spoiled GRE sequence allowing for a T1 weighting (T1W) and a balanced SSFP sequence allowing for a T2/T1W contrast were used for MR guidance. The contrast-to-noise-ratio (CNR) of the lesions was calculated and lesion conspicuity was assessed retrospectively (easily detectable / difficult to detect / not detectable). RESULTS: HCC was easily detectable in 33/52% (GRE/SSFP), difficult to detect in 30/18%, and not detectable in 37/30% of the cases. Mean CNR varied widely (9.1 for GRE vs. 16.4 for SSFP). Liver metastases were easily detectable in 58/41% (GRE/SSFP), difficult to detect in 14/21%, and not detectable in 28/38% of the cases. Mean CNR for liver metastases was 11.5 (GRE) vs. 12.7 (SSFP). Twenty percent of all lesions could not be detected with either of the MR fluoroscopy sequences. CONCLUSION: MR fluoroscopy using GRE and SSFP contrast enabled real-time detectability of 80% of the liver lesions.

Rapid freehand MR-guided percutaneous needle interventions: an image-based approach to improve workflow and feasibility.

PURPOSE: To develop and evaluate software-based methods for improving the workflow of magnetic resonance (MR)-guided percutaneous interventions. MATERIALS AND METHODS: A set of methods was developed that allows the user to: 1) plan an entire procedure, 2) directly apply this plan to skin entry site localization without further imaging, and 3) place a needle under real-time MR guidance with automatic alignment of three orthogonal slices along a planned trajectory with preference to the principal patient axes. To validate targeting accuracy and time, phantom experiments (96 targets) and in vivo paraspinal and kidney needle punctures in two pigs (55 targets) were performed. The influence of trajectory obliquity, level of experience, and organ motion on targeting accuracy and time was analyzed. RESULTS: Mean targeting error was 1.8 ± 0.9 mm (in vitro) and 2.9 ± 1.0 mm (in vivo) in all directions. No statistically significant differences in targeting accuracy between single- and double-oblique trajectories, novice and expert users, or paraspinal and kidney punctures were observed. The average time (in vivo) from trajectory planning to verification of accurate needle placement was 6 minutes. CONCLUSION: The developed methods allow for accurate needle placement along complex trajectories and are anticipated to reduce table time for MR-guided percutaneous needle interventions.

In vivo 31P MR spectroscopic imaging of the human prostate at 7 T: safety and feasibility.

(31)P MR spectroscopic imaging of the human prostate provides information about phosphorylated metabolites that could be used for prostate cancer characterization. The sensitivity of a magnetic field strength of 7 T might enable 3D (31)P MR spectroscopic imaging with relevant spatial resolution in a clinically acceptable measurement time. To this end, a (31)P endorectal coil was developed and combined with an eight-channel (1)H body-array coil to relate metabolic information to anatomical location. An extensive safety validation was performed to evaluate the specific absorption rate, the radiofrequency field distribution, and the temperature distribution of both coils. This validation consisted of detailed Finite Integration Technique simulations, confirmed by MR thermometry and B 1+ measurements in a phantom and in vivo temperature measurements. The safety studies demonstrated that the presence of the (31)P endorectal coil had no influence on the specific absorption rate levels and temperature distribution of the external eight-channel (1)H array coil. To stay within a 10 g averaged local specific absorption rate of 10 W/kg, a maximum time-averaged input power of 33 W for the (1)H array coil was allowed. For transmitting with the (31)P endorectal coil, our safety limit of less than 1°C temperature increase in vivo during a 15-min MR spectroscopic imaging experiment was reached at a time-averaged input power of 1.9 W. With this power setting, a second in vivo measurement was performed on a healthy volunteer. Using adiabatic excitation, 3D (31)P MR spectroscopic imaging produced spectra from the entire prostate in 18 min with a spatial resolution of 4 cm(3). The spectral resolution enabled the separate detection of phosphocholine, phosphoethanolamine, inorganic phosphate, and other metabolites that could play an important role in the characterization of prostate cancer.

Evaluation of a robotic technique for transrectal MRI-guided prostate biopsies.

OBJECTIVES: To evaluate the accuracy and speed of a novel robotic technique as an aid to perform magnetic resonance image (MRI)-guided prostate biopsies on patients with cancer suspicious regions. METHODS: A pneumatic controlled MR-compatible manipulator with 5 degrees of freedom was developed in-house to guide biopsies under real-time imaging. From 13 consecutive biopsy procedures, the targeting error, biopsy error and target displacement were calculated to evaluate the accuracy. The time was recorded to evaluate manipulation and procedure time. RESULTS: The robotic and manual techniques demonstrated comparable results regarding mean targeting error (5.7 vs 5.8 mm, respectively) and mean target displacement (6.6 vs 6.0 mm, respectively). The mean biopsy error was larger (6.5 vs 4.4 mm) when using the robotic technique, although not significant. Mean procedure and manipulation time were 76 min and 6 min, respectively using the robotic technique and 61 and 8 min with the manual technique. CONCLUSIONS: Although comparable results regarding accuracy and speed were found, the extended technical effort of the robotic technique make the manual technique - currently - more suitable to perform MRI-guided biopsies. Furthermore, this study provided a better insight in displacement of the target during in vivo biopsy procedures.