OUOrhan Unal

Orhan Unal

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Orhan Unal has a total of 79 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    AI-Powered Object Detection in Radiology: Current Models, Challenges, and Future Direction

    by Abdussalam Elhanashi, Sergio Saponara, Qinghe Zheng, Nawal Almutairi, Yashbir Singh, Shiba Kuanar, Farzana Ali, Orhan Unal, Shahriar Faghani

    Artificial intelligence (AI)-based object detection in radiology can assist in clinical diagnosis and treatment planning. This article examines the AI-based object detection models currently used in many imaging modalities, including X-ray Magnetic Resonance Imaging (MRI), Computed Tomography (CT), and Ultrasound (US). The key models from the convolutional neural network (CNN) as well as the contemporary transformer and hybrid models are analyzed based on their ability to detect pathological features, such as tumors, lesions, and tissue abnormalities. In addition, this review offers a closer look at the strengths and weaknesses of these models in terms of accuracy, robustness, and speed in real clinical settings. The common issues related to these models, including limited data, annotation quality, and interpretability of AI decisions, are discussed in detail. Moreover, the need for strong applicable models across different populations and imaging modalities are addressed. The importance of privacy and ethics in general data use as well as safety and regulations for healthcare data are emphasized. The future potential of these models lies in their accessibility in low resource settings, usability in shared learning spaces while maintaining privacy, and improvement in diagnostic accuracy through multimodal learning. This review also highlights the importance of interdisciplinary collaboration among artificial intelligence researchers, radiologists, and policymakers. Such cooperation is essential to address current challenges and to fully realize the potential of AI-based object detection in radiology.


  • Open Access

    Evaluation of a Deep Learning Reconstruction for High-Quality T2-Weighted Breast Magnetic Resonance Imaging

    by Timothy J. Allen, Leah C. Henze Bancroft, Orhan Unal, Lloyd D. Estkowski, Ty A. Cashen, Frank Korosec, Roberta M. Strigel, Frederick Kelcz, Amy M. Fowler, Alison Gegios, et al.

    Deep learning (DL) reconstruction techniques to improve MR image quality are becoming commercially available with the hope that they will be applicable to multiple imaging application sites and acquisition protocols. However, before clinical implementation, these methods must be validated for specific use cases. In this work, the quality of standard-of-care (SOC) T2w and a high-spatial-resolution (HR) imaging of the breast were assessed both with and without prototype DL reconstruction. Studies were performed using data collected from phantoms, 20 retrospectively collected SOC patient exams, and 56 prospectively acquired SOC and HR patient exams. Image quality was quantitatively assessed via signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), and edge sharpness. Qualitatively, all in vivo images were scored by either two or four radiologist readers using 5-point Likert scales in the following categories: artifacts, perceived sharpness, perceived SNR, and overall quality. Differences in reader scores were tested for significance. Reader preference and perception of signal intensity changes were also assessed. Application of the DL resulted in higher average SNR (1.2–2.8 times), CNR (1.0–1.8 times), and image sharpness (1.2–1.7 times). Qualitatively, the SOC acquisition with DL resulted in significantly improved image quality scores in all categories compared to non-DL images. HR acquisition with DL significantly increased SNR, sharpness, and overall quality compared to both the non-DL SOC and the non-DL HR images. The acquisition time for the HR data only required a 20% increase compared to the SOC acquisition and readers typically preferred DL images over non-DL counterparts. Overall, the DL reconstruction demonstrated improved T2w image quality in clinical breast MRI.


  • Open Access

    Automated Placement of Scan and Pre-Scan Volumes for Breast MRI Using a Convolutional Neural Network

    by Timothy J. Allen, Leah C. Henze Bancroft, Kang Wang, Ping Ni Wang, Orhan Unal, Lloyd D. Estkowski, Ty A. Cashen, Ersin Bayram, Roberta M. Strigel, James H. Holmes

    Graphically prescribed patient-specific imaging volumes and local pre-scan volumes are routinely placed by MRI technologists to optimize image quality. However, manual placement of these volumes by MR technologists is time-consuming, tedious, and subject to intra- and inter-operator variability. Resolving these bottlenecks is critical with the rise in abbreviated breast MRI exams for screening purposes. This work proposes an automated approach for the placement of scan and pre-scan volumes for breast MRI. Anatomic 3-plane scout image series and associated scan volumes were retrospectively collected from 333 clinical breast exams acquired on 10 individual MRI scanners. Bilateral pre-scan volumes were also generated and reviewed in consensus by three MR physicists. A deep convolutional neural network was trained to predict both the scan and pre-scan volumes from the 3-plane scout images. The agreement between the network-predicted volumes and the clinical scan volumes or physicist-placed pre-scan volumes was evaluated using the intersection over union, the absolute distance between volume centers, and the difference in volume sizes. The scan volume model achieved a median 3D intersection over union of 0.69. The median error in scan volume location was 2.7 cm and the median size error was 2%. The median 3D intersection over union for the pre-scan placement was 0.68 with no significant difference in mean value between the left and right pre-scan volumes. The median error in the pre-scan volume location was 1.3 cm and the median size error was −2%. The average estimated uncertainty in positioning or volume size for both models ranged from 0.2 to 3.4 cm. Overall, this work demonstrates the feasibility of an automated approach for the placement of scan and pre-scan volumes based on a neural network model.


  • Open Access

    Value CMR: Towards a Comprehensive, Rapid, Cost-Effective Cardiovascular Magnetic Resonance Imaging

    by El-Sayed H. Ibrahim, Luba Frank, Dhiraj Baruah, V. Emre Arpinar, Andrew S. Nencka, Kevin M. Koch, L. Tugan Muftuler, Orhan Unal, Jadranka Stojanovska, Jason C. Rubenstein, et al.

    Cardiac magnetic resonance imaging (CMR) is considered the gold standard for measuring cardiac function. Further, in a single CMR exam, information about cardiac structure, tissue composition, and blood flow could be obtained. Nevertheless, CMR is underutilized due to long scanning times, the need for multiple breath-holds, use of a contrast agent, and relatively high cost. In this work, we propose a rapid, comprehensive, contrast-free CMR exam that does not require repeated breath-holds, based on recent developments in imaging sequences. Time-consuming conventional sequences have been replaced by advanced sequences in the proposed CMR exam. Specifically, conventional 2D cine and phase-contrast (PC) sequences have been replaced by optimized 3D-cine and 4D-flow sequences, respectively. Furthermore, conventional myocardial tagging has been replaced by fast strain-encoding (SENC) imaging. Finally, T1 and T2 mapping sequences are included in the proposed exam, which allows for myocardial tissue characterization. The proposed rapid exam has been tested in vivo. The proposed exam reduced the scan time from >1 hour with conventional sequences to <20 minutes. Corresponding cardiovascular measurements from the proposed rapid CMR exam showed good agreement with those from conventional sequences and showed that they can differentiate between healthy volunteers and patients. Compared to 2D cine imaging that requires 12-16 separate breath-holds, the implemented 3D-cine sequence allows for whole heart coverage in 1-2 breath-holds. The 4D-flow sequence allows for whole-chest coverage in less than 10 minutes. Finally, SENC imaging reduces scan time to only one slice per heartbeat. In conclusion, the proposed rapid, contrast-free, and comprehensive cardiovascular exam does not require repeated breath-holds or to be supervised by a cardiac imager. These improvements make it tolerable by patients and would help improve cost effectiveness of CMR and increase its adoption in clinical practice.


  • Open Access

    Value CMR: Towards a Comprehensive, Rapid, Cost-Effective Cardiovascular Magnetic Resonance Imaging

    by El-Sayed H. Ibrahim, Luba Frank, Dhiraj Baruah, Jason C. Rubenstein, V. Emre Arpinar, Andrew S. Nencka, Kevin M. Koch, L Tugan Muftuler, Orhan Unal, Jadranka Stojanovska, et al.

    CMR is considered the gold standard for measuring heart function, including cardiac volumes and mass. Further, in a single CMR exam, information about cardiac function, structure, tissue composition, and blood flow could be obtained. Nevertheless, CMR is underutilized due to long scanning times, the need for multiple breath-holds, use of a contrast agent, and relatively higher cost compared to echocardiography. In this study, we propose a rapid CMR exam based on recent developments in imaging sequences. The proposed exam is both rapid and provides comprehensive cardiovascular information without the need for a contrast agent or multiple breath-holds. The developed exam includes advanced sequences for evaluating global and regional cardiac functions, myocardial tissue characterization, and flow hemodynamics in the heart, valves, and large vessels. Time-consuming conventional sequences have been replaced by advanced sequences, which resulted in reducing scan time from > 1 hour with conventional CMR exam to <20 minutes with the proposed rapid CMR exam. Specifically, conventional two-dimensional (2D) cine and phase-contrast (PC) sequences have been replaced by optimized three-dimensional (3D)-cine and four-dimensional (4D)-flow sequences, respectively. Compared to 2D cine imaging that requires 12-16 separate breath-holds, the implemented 3D-cine sequence allows for whole heart coverage in 1-2 breath-holds; thus, reducing scan time by 80-90%. Similarly, compared to 2D PC flow imaging that requires multiple breath-holds and the presence of an experienced cardiac operator for precise prescription of the imaging planes, the implemented 4D-flow sequence allows for whole-chest coverage in ∼10-minute, free-breathing acquisition without the need for a navigator echo, which makes scan time independent of the patient’s breathing pattern. Furthermore, conventional myocardial tagging has been replaced by the fast strain-encoding (SENC) sequence, which reduces scan time from one slice per breath-hold to only one slice per heartbeat (∼1 second). Finally, T1 and T2 mapping sequences are included in this proposed exam, which allow for myocardial tissue characterization without the need for contrast. The proposed rapid exam has been tested on volunteers and measurements showed good agreement with those from conventional sequences despite the significant reduction in scan time. In conclusion, we propose a rapid, contrast-free, and comprehensive cardiovascular exam that does not require repeated breath-holds or a cardiac experienced operator to run the exam, which would result in improving cost effectiveness of CMR and increasing its adoption in clinical practice.


  • Open Access

    A multi-vendor, multi-center study on reproducibility and comparability of fast strain-encoded cardiovascular magnetic resonance imaging

    by Jennifer Erley, Victoria Zieschang, Tomas Lapinskas, Aylin Demir, Stephanie Wiesemann, Markus Haass, Nael F Osman, Orlando P Simonetti, Yingmin Liu, Amit R Patel, et al.

    Myocardial strain is a convenient parameter to quantify left ventricular (LV) function. Fast strain-encoding (fSENC) enables the acquisition of cardiovascular magnetic resonance images for strain-measurement within a few heartbeats during free-breathing. It is necessary to analyze inter-vendor agreement of techniques to determine strain, such as fSENC, in order to compare existing studies and plan multi-center studies. Therefore, the aim of this study was to investigate inter-vendor agreement and test-retest reproducibility of fSENC for three major MRI-vendors. fSENC-images were acquired three times in the same group of 15 healthy volunteers using 3 Tesla scanners from three different vendors: at the German Heart Institute Berlin, the Charité University Medicine Berlin-Campus Buch and the Theresien-Hospital Mannheim. Volunteers were scanned using the same imaging protocol composed of two fSENC-acquisitions, a 15-min break and another two fSENC-acquisitions. LV global longitudinal and circumferential strain (GLS, GCS) were analyzed by a trained observer (Myostrain 5.0, Myocardial Solutions) and for nine volunteers repeatedly by another observer. Inter-vendor agreement was determined using Bland-Altman analysis. Test-retest reproducibility and intra- and inter-observer reproducibility were analyzed using intraclass correlation coefficient (ICC) and coefficients of variation (CoV). Inter-vendor agreement between all three sites was good for GLS and GCS, with biases of 0.01–1.88%. Test-retest reproducibility of scans before and after the break was high, shown by ICC- and CoV values of 0.63–0.97 and 3–9% for GLS and 0.69–0.82 and 4–7% for GCS, respectively. Intra- and inter-observer reproducibility were excellent for both parameters (ICC of 0.77–0.99, CoV of 2–5%). This trial demonstrates good inter-vendor agreement and test–retest reproducibility of GLS and GCS measurements, acquired at three different scanners from three different vendors using fSENC. The results indicate that it is necessary to account for a possible bias (< 2%) when comparing strain measurements of different scanners. Technical differences between scanners, which impact inter-vendor agreement, should be further analyzed and minimized.DRKS Registration Number: 00013253.Universal Trial Number (UTN): U1111-1207-5874.


  • Open Access

    Treatment Results of Patients Followed in Intensive Care Unit in Severe Maternal Morbidity Cases

    by Hilal Uslu Yuvaci, Tuba Duzcan, Nermin Akdemir, Erman Sever, Selcuk Ozden, Arif Serhan Cevrioglu, Orhan Unal, Funda Tozlu, Havva Sayhan

    Objective: The aim of this study was the evaluation of the characteristics and treatment results of women, who were brought to the intensive care unit due to severe obstetric morbidity, together with severe and acute complications in pregnancy. Study Design: Pregnant and puerperal women treated in the intensive care unit of a tertiary care center during two years of time were included in this study, and their files were scanned retrospectively. Patients’ demographic characteristics, length of stay, point of entry, neonatal results, birth statuses and modes of delivery, hemodynamic data, and histories were recorded. The qualitative data were evaluated by the Pearson Chi-squared, Fisher Freeman Halton, and Fisher’s Exact tests. Results: In the two-year period, 16,728 births occurred at our hospital. 68 cases among them were accompanied with severe maternal morbidity, and 2 maternal deaths were observed in our clinic. Of the cases, 58.8% (n=40) had severe hypertensive diseases, whereas 35.3% of the cases (n=24) had obstetric complications that developed due to bleeding. While 40% of the cases (n=16) with hypertension had severe pre-eclampsia, 35% of the cases (n=14) had eclampsia and 25% (n=10) had HELLP. Conclusions: The most important reasons for severe maternal morbidity are the complications related to obstetric bleeding and hypertensive diseases related to pregnancy. Early diagnosis of the obstetric complication risk factors is necessary for preventing maternal morbidity. Antenatal follow-ups and the births of high risk pregnancy patients should be performed in tertiary centers.


  • Modeling Endovascular MRI Coil Coupling With Transmit RF Excitation

    by Madhav Venkateswaran, Orhan Unal, Samuel Hurley, Alexey Samsonov, Peng Wang, Sean B. Fain, Krishna N. Kurpad

    Objective: To model inductive coupling of endovascular coils with transmit RF excitation for selecting coils for MRI-guided interventions. Methods: Independent and computationally efficient FEM models are developed for the endovascular coil, cable, transmit excitation, and imaging domain. Electromagnetic and circuit solvers are coupled to simulate net B1 + fields and induced currents and voltages. Our models are validated using the Bloch-Siegert B1 + mapping sequence for a series-tuned multimode coil, capable of tracking, wireless visualization, and high-resolution endovascular imaging. Results: Validation shows good agreement at 24-, 28-, and 34-μT background RF excitation within experimental limitations. Quantitative coil performance metrics agree with simulation. A parametric study demonstrates tradeoff in coil performance metrics when varying number of coil turns. Tracking, imaging, and wireless marker multimode coil features and their integration is demonstrated in a pig study. Conclusion: Developed models for the multimode coil were successfully validated. Modeling for geometric optimization and coil selection serves as a precursor to time consuming and expensive experiments. Specific applications demonstrated include parametric optimization, coil selection for a cardiac intervention, and an animal imaging experiment. Significance: Our modular, adaptable, and computationally efficient modeling approach enables rapid comparison, selection, and optimization of inductively coupled coils for MRI-guided interventions.


  • Open Access

    Volumetric late gadolinium‐enhanced myocardial imaging with retrospective inversion time selection

    by Steve Kecskemeti, Kevin Johnson, Christopher J. François, Mark L. Schiebler, Orhan Unal

    Purpose To develop and validate a novel free‐breathing 3D radial late gadolinium‐enhanced magnetic resonance imaging technique (3D LGE‐MRI) with isotropic resolution and retrospective inversion time (TI) selection for myocardial viability imaging. Materials and Methods The 3D radial LGE‐MRI method featuring an interleaved and bit‐reversed radial k‐space trajectory was evaluated in 12 subjects that also had clinical breath‐hold Cartesian 2D LGE‐MRI. The 3D LGE‐MRI acquisition requires a predicted TI and a user‐controlled data acquisition window that determines the sampling width around the predicted TI. Sliding window reconstructions with update rates of 1× the repetition time (TR) allow for a user selectable TI to obtain the maximum nulling of the myocardium. The retrospective nature of the acquisition allows the user to choose from a range of possible TI times centered on the expected TI. Those projections most corrupted by respiratory motion, as determined by a respiratory bellows signal, were resampled according to the diminishing variance algorithm. The quality of the left ventricular myocardial nulling on the 3D LGE‐MRI and 2D LGE‐MRI was assessed using a 4‐point Likert scale by two experienced radiologists. Comparison of image quality scores for the two methods was performed using generalized estimating equations. Results All 3D LGE‐MRI cases produced similar nulling of myocardial signal as the 2D LGE‐MRI. The image quality of myocardial nulling was not significantly different between the two acquisitions (mean nulling of 3.4 for 2D vs. 3.1 for 3D, and P = 0.0645). The average absolute deviation from mean scores was also not determined to be statistically significant (1.8 for 2D and 0.4 for 3D and P = 0.1673). Total acquisition time was ∼9 minutes for 3D LGE‐MRI with voxel sizes ranging from 1.63 to 2.03 mm3. Conversely, the total imaging time was twice as long for the 2D DCE‐MRI (>17 minutes) with an eight times larger voxel size of 1.4 × 2.2 × 7.0 mm. Conclusion The 3D LGE‐MRI technique demonstrated in this study is a promising alternative for the assessment of myocardial viability in patients who have difficulty sustaining breath‐holds for the clinical standard 2D LGE‐MRI. J. Magn. Reson. Imaging 2013;38:1276–1282.


  • Open Access

    MR‐visible coatings for endovascular device visualization

    by Orhan Unal, Junwei Li, Weiguo Cheng, Hyuk Yu, Charles M. Strother

    Purpose To investigate the potential utility of magnetic resonance (MR)‐visible coatings for passive visualization of therapeutic endovascular devices such as catheters and guidewires. Materials and Methods Using a multistep coating process, gadolinium‐based coatings were applied to commercially available off‐the‐shelf catheters and guidewires. These coated devices were imaged in phantoms made of fat‐free yogurt, saline, and whole blood and also in live canine aorta on a 1.5‐T cardiovascular MR scanner using T1‐weighted two‐dimensional radiofrequency (RF)‐spoiled gradient‐recalled echo, two‐dimensional spin echo, and three‐dimensional RF‐spoiled gradient‐recalled echo techniques. Results Commercially available off‐the shelf catheters (4, 5, and 6 French) and guidewires (0.038 inch) were clearly visualized in all phantoms and canine aorta and the coatings proved to be durable and imageable without degradation in signal intensity up to 24 hours. MR‐visible coatings address some of the shortcomings that have previously limited the role of MR as a guidance tool. Conclusion Both in vitro and in vivo visualization of therapeutic endovascular devices coated with MR‐visible coatings are found to be clinically viable. J. Magn. Reson. Imaging 2006.