Addition of 131I-MIBG to PRRT (90Y-DOTATOC) for Personalized Treatment of Selected Patients with Neuroendocrine Tumors.

Peptide receptor radionuclide therapy (PRRT) is an effective treatment for metastatic neuroendocrine tumors. Delivering a sufficient tumor radiation dose remains challenging because of critical-organ dose limitations. Adding 131I-metaiodobenzylguanidine (131I-MIBG) to PRRT may be advantageous in this regard. Methods: A phase 1 clinical trial was initiated for patients with nonoperable progressive neuroendocrine tumors using a combination of 90Y-DOTATOC plus 131I-MIBG. Treatment cohorts were defined by radiation dose limits to the kidneys and the bone marrow. Subject-specific dosimetry was used to determine the administered activity levels. Results: The first cohort treated subjects to a dose limit of 1,900 cGy to the kidneys and 150 cGy to the marrow. No dose-limiting toxicities were observed. Tumor dosimetry estimates demonstrated an expected dose increase of 34%-83% using combination therapy as opposed to 90Y-DOTATOC PRRT alone. Conclusion: These findings demonstrate the feasibility of using organ dose for a phase 1 escalation design and suggest the safety of using 90Y-DOTATOC and 131I-MIBG.

90Y-DOTATOC Dosimetry-Based Personalized Peptide Receptor Radionuclide Therapy.

Pretherapy PET with 86Y-DOTATOC is considered the ideal dosimetry protocol for 90Y-DOTATOC therapy; however, its cost, limited availability, and need for infusion of amino acids to mimic the therapy administration limit its use in the clinical setting. The goal of this study was to develop a dosimetric method for 90Y-DOTATOC using 90Y-DOTATOC PET/CT and bremsstrahlung SPECT/CT and to determine whether dosimetry-based administered activities differ significantly from standard administered activities. Methods: This was a prospective phase 2 trial of 90Y-DOTATOC therapy in patients with somatostatin receptor-positive tumors. 90Y-DOTATOC was given in 3 cycles 6-8 wk apart. In the first cycle of therapy, adults received 4.4 GBq and children received 1.85 GBq/m2; the subsequent administered activities were adjusted according to the dosimetry of the preceding cycle so as not to exceed a total kidney dose of 23 Gy and bone marrow dose of 2 Gy. The radiation dose to the kidneys was determined from serial imaging sessions consisting of time-of-flight 90Y-DOTATOC PET/CT at 5 h after therapy and 90Y-DOTATOC bremsstrahlung SPECT/CT at 6, 24, 48, and 72 h. The PET/CT data were used to measure the absolute concentration of 90Y-DOTATOC and to calibrate the bremsstrahlung SPECT kidney clearance data. The radiation dose to the kidneys was determined by multiplying the time-integrated activity (from the fitted biexponential curve of renal clearance of 90Y-DOTATOC) with the energy emitted per decay, divided by the mass of the kidneys. Results: The radiation dose to the kidneys per cycle of 90Y-DOTATOC therapy was highly variable among patients, ranging from 0.32 to 3.0 mGy/MBq. In 17 (85%) of the 20 adult patients who received the second and the third treatment cycles of 90Y-DOTATOC, the administered activity was modified by at least 20% from the starting administered activity. Conclusion: Renal dosimetry of 90Y-DOTATOC is feasible using 90Y-DOTATOC time-of-flight PET/CT and bremsstrahlung SPECT/CT and has a significant impact on the administered activity in treatment cycles.

Technical Note: Single time point dose estimate for exponential clearance.

OBJECTIVE: Although personalized dosimetry may be desirable for radionuclide therapy treatments, the multiple time samples required to determine the total integrated activity puts a burden on patients and clinic resources. The aim of this paper is to demonstrate that when some prior knowledge is known about the tracer kinetic parameters, the total integrated activity (and thus radiation dose) can be estimated from a single time sample. METHODS: Mathematical derivations have been performed to generate equations for the total integrated activity in terms of a single time sample of activity for monoexponential and biexponential clearance. Simulations were performed using both exponential models where the rate constants and associated parameters were randomly sampled from distributions with a known mean. The actual total integrated activity for each random sample was compared with the estimated total integrated activity using the mean value of the parameters. Retrospective analysis of 90 Y DOTATOC data from a clinical trial provided a comparison of actual kidney dose with the estimated kidney dose using the single time point approach. RESULTS: The optimal sampling time for the single point approach was found to be equal to the mean time of the rate constant. The simulation results for the monoexponential and biexpoential models were similar. Regressions comparing the actual and estimated total integrated activity had very high correlations (r2  > 0.95) along with acceptable standard errors of estimate, especially at the optimal sampling point. The retrospective analysis of the 90 Y DOTATOC data also yielded similar results with an r2  = 0.95 and a standard error of estimate of 61 cGy. CONCLUSIONS: In situations where there is prior knowledge about the population averages of kinetic parameters, these results suggest that the single time point approach can be used to estimate the total integrated activity and dose with sufficient accuracy to manage radionuclide therapy. This will make personalized dosimetry much easier to perform and more available to the community.

Effect of fatigue on reading computed tomography examination of the multiply injured patient.

Our goal was to ascertain how fatigue affects performance in reading computed tomography (CT) examinations of patients with multiple injuries. CT images with multiple fractures from a previous study of satisfaction of search (SOS) were read by radiologists after a day of clinical work. Performance in this study with fatigued readers was compared to a previous study in which readers were not fatigued. Detection accuracy for obvious injuries was not affected by fatigue, but accuracy for subtle fractures was reduced ([Formula: see text]). An SOS effect on decision thresholds was evident mirroring recent studies. Without fatigue, readers spent more time interpreting and reporting findings as the number of the injuries increased. When fatigued, readers did not increase reading time as fracture number increased. Without fractures, reading time for not-fatigued and fatigued readers was the same ([Formula: see text]) but was significant ([Formula: see text]) with an added subtle fracture. The difference increased with a major injury ([Formula: see text]) and increased further with both a major injury and subtle fracture ([Formula: see text]). Fatigue and multiple abnormalities have independent effects on detection performance but do interact in determining search time.

The Impact of Fatigue on Satisfaction of Search in Chest Radiography.

RATIONALE AND OBJECTIVES: To assess the nature of the satisfaction of search (SOS) effect in chest radiography when observers are fatigued; determine if we could replicate recent findings that have documented the nature of the SOS effect to be due to a threshold shift rather than a change in diagnostic accuracy as in earlier film-based studies. MATERIALS AND METHODS: Nearing or at the end of a clinical workday, 20 radiologists read 64 chest images twice, once with and once without the addition of a simulated pulmonary nodule. Half of the images had different types of "test" abnormalities. Decision thresholds were analyzed using the center of the range of false-positive (FP) and true-positive (TP) fractions associated with each receiver operating characteristic (ROC) point for reporting test abnormalities. Detection accuracy was assessed with ROC technique and inspection time was recorded. RESULTS: The SOS effect was confirmed to be a reduction in willingness to respond (threshold shift). The center of the FP range was significantly reduced (FP = 0.10 without added nodules, FP = 0.05 with added nodules, F(1,18) = 19.85, P = 0.0003). The center of the TP range was significantly reduced (TP = 0.39 without added nodules, TP = 0.33 with added nodules, F(1,18) = 10.81, P = 0.004). CONCLUSIONS: This study suggests that fatigue does not change the nature of the SOS effect, but rather may be additive with the SOS effect. SOS reduces both TP and FP responses, whereas fatigue reduces TPs more than FPs.

Trauma in CT: The Role of Severe Injury on Satisfaction of Search Revised.

PURPOSE: The satisfaction-of-search (SOS) effect occurs when an abnormality on an image is missed because another is found. The aim of this experiment was to test whether severe distracting fractures control the magnitude of SOS on other fractures when both appear in a single CT image. METHODS: The institutional review board approved this study. The experimental (SOS) condition included 35 cervical spine CT cases, all of which contained severe cervical spine injuries. For each of these cases, a similar case was found that had no injuries. Image modification software was developed to add simulated fractures to each pair of cases, with and without a major injury. Sixteen different minor fractures were added to 16 of the 35 pairs of images. The 35 cases without native injuries constituted a control (non-SOS) condition mixed in a random order. Twenty radiologists read 35 mixed cases in each of two sessions. False-positive evaluations were collected only for cases without simulated fractures. RESULTS: An SOS effect on the detection of simulated fractures was not observed. There was a nonsignificant (P = .07) finding of poorer detection in the presence of cases with severe injuries. However, the magnitude of the effect was no greater than has been observed for less severe distracting injuries. CONCLUSIONS: The outcome agrees with the results of two previous experiments that failed to yield an SOS effect associated with detecting severe injuries, suggesting that the severity of a distracting injury does not determine whether a second injury is discovered.

The Influence of a Vocalized Checklist on Detection of Multiple Abnormalities in Chest Radiography.

RATIONALE AND OBJECTIVES: Although a checklist has been recommended for preventing satisfaction of search (SOS) errors, a previous research study did not demonstrate that benefit. However, observers in that study had to turn away from the image display to use the checklist. The current study tested a vocalized checklist to avoid this constraint. MATERIALS AND METHODS: A total of 64 chest computed radiographs, half containing various "test" abnormalities, were read twice by 20 radiologists, once with and once without the addition of a simulated pulmonary nodule. Readers used a vocalized checklist-directing search. Receiver operating characteristic (ROC) detection accuracy and decision thresholds were analyzed to study the effects of adding the nodule on detecting the test abnormalities. RESULTS: Adding nodules induced a substantial reluctance to report the other abnormalities (P < 0.001), as had been the case in the most recent study of the SOS effect in radiography. CONCLUSIONS: The vocalized checklist did not reduce nor eliminate the SOS effect on readiness to report further abnormalities. Although useful for organizing search and reporting, particularly among students, a vocalized checklist does not prevent SOS effects.

Satisfaction of Search in Chest Radiography 2015.

RATIONALE AND OBJECTIVES: Two decades have passed since the publication of laboratory studies of satisfaction of search (SOS) in chest radiography. Those studies were performed using film. The current investigation tests for SOS effects in computed radiography of the chest. METHODS: Sixty-four chest computed radiographs half demonstrating various "test" abnormalities were read twice by 20 radiologists, once with and once without the addition of a simulated pulmonary nodule. Receiver-operating characteristic detection accuracy and decision thresholds were analyzed to study the effects of adding the nodule on detecting the test abnormalities. Results of previous studies were reanalyzed using similar modern techniques. RESULTS: In the present study, adding nodules did not influence detection accuracy for the other abnormalities (P = .93), but did induce a reluctance to report them (P < .001). Adding nodules did not affect inspection time (P = .58) so the reluctance to report was not associated with reduced search. Reanalysis revealed a similar decision threshold shift that had not been recognized in the early studies of SOS in chest radiography (P < .01) in addition to reduced detection accuracy (P < .01). CONCLUSIONS: The nature of SOS in chest radiography has changed, but it is not clear why. ADVANCES IN KNOWLEDGE: SOS may be changing as a function of changes in radiology education and practice.

Improved nuclear medicine uniformity assessment with noise texture analysis.

UNLABELLED: Because γ cameras are generally susceptible to environmental conditions and system vulnerabilities, they require routine evaluation of uniformity performance. The metrics for such evaluations are commonly pixel value-based. Although these metrics are typically successful at identifying regional nonuniformities, they often do not adequately reflect subtle periodic structures; therefore, additional visual inspections are required. The goal of this project was to develop, test, and validate a new uniformity analysis metric capable of accurately identifying structures and patterns present in nuclear medicine flood-field uniformity images. METHODS: A new uniformity assessment metric, termed the structured noise index (SNI), was based on the 2-dimensional noise power spectrum (NPS). The contribution of quantum noise was subtracted from the NPS of a flood-field uniformity image, resulting in an NPS representing image artifacts. A visual response filter function was then applied to both the original NPS and the artifact NPS. A single quantitative score was calculated on the basis of the magnitude of the artifact. To verify the validity of the SNI, an observer study was performed with 5 expert nuclear medicine physicists. The correlation between the SNI and the visual score was assessed with Spearman rank correlation analysis. The SNI was also compared with pixel value-based assessment metrics modeled on the National Electrical Manufacturers Association standard for integral uniformity in both the useful field of view (UFOV) and the central field of view (CFOV). RESULTS: The SNI outperformed the pixel value-based metrics in terms of its correlation with the visual score (ρ values for the SNI, integral UFOV, and integral CFOV were 0.86, 0.59, and 0.58, respectively). The SNI had 100% sensitivity for identifying both structured and nonstructured nonuniformities; for the integral UFOV and CFOV metrics, the sensitivities were only 62% and 54%, respectively. The overall positive predictive value of the SNI was 87%; for the integral UFOV and CFOV metrics, the positive predictive values were only 67% and 50%, respectively. CONCLUSION: The SNI accurately identified both structured and nonstructured flood-field nonuniformities and correlated closely with expert visual assessment. Compared with traditional pixel value-based analysis, the SNI showed superior performance in terms of its correlation with visual perception. The SNI method is effective for detecting and quantifying visually apparent nonuniformities and may reduce the need for more subjective visual analyses.

Feasibility and advantage of adding (131)I-MIBG to (90)Y-DOTATOC for treatment of patients with advanced stage neuroendocrine tumors.

BACKGROUND: Peptide receptor radionuclide therapy (PRRT) is an effective form of treatment for patients with metastatic neuroendocrine tumors (NETs). However, delivering sufficient radiation dose to the tumor to result in a high percentage of long-term tumor remissions remains challenging because of the limits imposed on administered activity levels by radiation damage to normal tissues. The goal of this study was to evaluate the dosimetric advantages of adding (131)I meta-iodobenzylguanidine ((131)I-MIBG) to (90)Y DOTA Phe1-Tyr3-octreotide ((90)Y-DOTATOC) in patients with advanced stage midgut NETs. METHODS: Ten patients were imaged simultaneously with (131)I-MIBG and (111)In-pentetreotide (as a surrogate for (90)Y-DOTATOC) on days 1, 2, and 3 post-administration. Blood samples were obtained at the same time points. Using dosimetry measures from this data and our previously published methodology for calculating optimal combined administered activity levels for therapy, we determined the amount of (131)I-MIBG that could be added to (90)Y-DOTATOC without exceeding normal organ dose limits (marrow and kidneys) along with the expected increase in associated tumor dose, if any. RESULTS: We found that a median value of 34.6 GBq of (131)I-MIBG could be safely added to (90)Y-DOTATOC (delivered over multiple cycles) by reducing the maximum total deliverable (90)Y-DOTATOC by a median value of 24.5%. Taking this treatment approach, we found that there would be a median increase in deliverable tumor dose of 4,046 cGy in six of the ten subjects. Of note, there were a small number of metastases that were positive for only one or the other of these radiopharmaceuticals within the same subject. CONCLUSIONS: We conclude that approximately half of the patients with midgut NETs that are eligible for PRRT could reasonably be expected to benefit from the addition of (131)I-MIBG to (90)Y-DOTATOC.

Repeatability of gallium-68 DOTATOC positron emission tomographic imaging in neuroendocrine tumors.

OBJECTIVE: To evaluate the repeatability of gallium-68 1,4,7,10-tetraazacyclododecane-N,N',N'',N'''-tetraacetic (DOTA)-D-Phe1-Try3-octreotide (68Ga-DOTATOC) positron emission tomography (PET) in neuroendocrine tumors. METHODS: Five patients with neuroendocrine tumors were imaged with 68Ga-DOTATOC PET twice within 5 days. Maximum and mean standardized uptake values (SUVmax and SUVmean) and kinetic parameters (K-Patlak and K-influx) of target lesions were measured. The repeatability of these measurements was investigated. RESULTS: Forty-seven target lesions were identified on whole-body PET and 21 lesions on dynamic images. There was excellent repeatability with intraclass correlation coefficient of 0.99 for SUVmax, SUVmean, and K-Patlak, and 0.85 for K-influx. The median absolute percent differences and the interquartile ranges (IQR) between 2 scans for SUVmax and SUVmean were 7.4% (IQR, 14.1%) and 9.3% (IQR, 10.6%), respectively. The median absolute percent differences for K-Patlak and K-influx were 12.5% (IQR, 12.6%) and 29.9% (IQR, 22.4%), respectively. The SUVmax of target lesions did not differ by more than 25% between the 2 scans. CONCLUSIONS: 68Ga-DOTATOC PET imaging of neuroendocrine tumors is highly reproducible. A difference of more than 25% in SUVmax represents a change that is larger than the measurement error observed on repeated studies and should reflect a significant change in the biological character of the tumor.

Satisfaction of search from detection of pulmonary nodules in computed tomography of the chest.

RATIONALE AND OBJECTIVES: We tested whether satisfaction of search (SOS) effects that occur in computed tomography (CT) examination of the chest on detection of native abnormalities are produced by the addition of simulated pulmonary nodules. MATERIALS AND METHODS: Two experiments were conducted. In the first experiment, 70 CT examinations, half that demonstrated diverse, subtle abnormalities and half that demonstrated no native lesions, were read by 18 radiology residents and fellows under two experimental conditions: presented with and without pulmonary nodules. In a second experiment, many of the examinations were replaced to include more salient native abnormalities. This set was read by 14 additional radiology residents and fellows. In both experiments, detection of the natural abnormalities was studied. Receiver operating characteristic (ROC) curve areas for each reader-treatment combination were estimated using empirical and proper ROC models. Additional analyses focused on decision thresholds and visual search time on abnormality-free CT slice ranges. Institutional review board approval and informed consent from 32 participants were obtained. RESULTS: Observers more often missed diverse native abnormalities when pulmonary nodules were added, but also made fewer false-positive responses. There was no change in ROC area, but decision criteria grew more conservative. The SOS effect on decision thresholds was accompanied by a reduction in search time on abnormality-free CT slice ranges. CONCLUSION: The SOS effect in CT examination of the chest is similar to that found in contrast examination of the abdomen, involving induced visual neglect.

Do long radiology workdays affect nodule detection in dynamic CT interpretation?

PURPOSE: A previous study demonstrated decreased diagnostic accuracy for finding fractures and decreased ability to focus on skeletal radiographs after a long working day. Skeletal radiographic examinations commonly have images that are displayed statically. The aim of this study was to investigate whether diagnostic accuracy for detecting pulmonary nodules on CT of the chest displayed dynamically would be similarly affected by fatigue. METHODS: Twenty-two radiologists and 22 residents were given 2 tests searching CT chest sequences for a solitary pulmonary nodule before and after a day of clinical reading. To measure search time, 10 lung CT sequences, each containing 20 consecutive sections and a single nodule, were inspected using free search and navigation. To measure diagnostic accuracy, 100 CT sequences, each with 20 sections and half with nodules, were displayed at preset scrolling speed and duration. Accuracy was measured using receiver operating characteristic curve analysis. Visual strain was measured via dark vergence, an indicator of the ability to keep the eyes focused on the display. RESULTS: Diagnostic accuracy was reduced after a day of clinical reading (P = .0246), but search time was not affected (P > .05). After a day of reading, dark vergence was significantly larger and more variable (P = .0098), reflecting higher levels of visual strain, and subjective ratings of fatigue were also higher. CONCLUSIONS: After their usual workday, radiologists experience increased fatigue and decreased diagnostic accuracy for detecting pulmonary nodules on CT. Effects of fatigue may be mitigated by active interaction with the display.

Toward fully automated processing of dynamic susceptibility contrast perfusion MRI for acute ischemic cerebral stroke.

We developed fully automated software for dynamic susceptibility contrast (DSC) MR perfusion-weighted imaging (PWI) to efficiently and reliably derive critical hemodynamic information for acute stroke treatment decisions. Brain MR PWI was performed in 80 consecutive patients with acute nonlacunar ischemic stroke within 24h after onset of symptom from January 2008 to August 2009. These studies were automatically processed to generate hemodynamic parameters that included cerebral blood flow and cerebral blood volume, and the mean transit time (MTT). To develop reliable software for PWI analysis, we used computationally robust algorithms including the piecewise continuous regression method to determine bolus arrival time (BAT), log-linear curve fitting, arrival time independent deconvolution method and sophisticated motion correction methods. An optimal arterial input function (AIF) search algorithm using a new artery-likelihood metric was also developed. Anatomical locations of the automatically determined AIF were reviewed and validated. The automatically computed BAT values were statistically compared with estimated BAT by a single observer. In addition, gamma-variate curve-fitting errors of AIF and inter-subject variability of AIFs were analyzed. Lastly, two observes independently assessed the quality and area of hypoperfusion mismatched with restricted diffusion area from motion corrected MTT maps and compared that with time-to-peak (TTP) maps using the standard approach. The AIF was identified within an arterial branch and enhanced areas of perfusion deficit were visualized in all evaluated cases. Total processing time was 10.9+/-2.5s (mean+/-s.d.) without motion correction and 267+/-80s (mean+/-s.d.) with motion correction on a standard personal computer. The MTT map produced with our software adequately estimated brain areas with perfusion deficit and was significantly less affected by random noise of the PWI when compared with the TTP map. Results of image quality assessment by two observers revealed that the MTT maps exhibited superior quality over the TTP maps (88% good rating of MTT as compared to 68% of TTP). Our software allowed fully automated deconvolution analysis of DSC PWI using proven efficient algorithms that can be applied to acute stroke treatment decisions. Our streamlined method also offers promise for further development of automated quantitative analysis of the ischemic penumbra.

Modeling combined radiopharmaceutical therapy: a linear optimization framework.

In this paper, we investigate a previously proposed mathematical model describing the effects that an innovative combined radiopharmaceutical therapy might have on the delivery of radiation to the tumor and limiting critical organs. While focused on a specific dual agent therapy, this investigation will prove mathematically that for any two therapeutic radiopharmaceuticals with different limiting critical organs the model provides patient specific conditions under which combination therapy is superior to single agent therapy. In addition, this paper outlines general methods for calculating the amounts of administered radioactivity for each drug required to optimize tumor radiation dose. We also consider extensions of this model to include an arbitrary number of independent radiopharmaceuticals and/or other treatment factors.

Recent advances in SPECT imaging.

SPECT is a rapidly changing field, and the past several years have produced new developments in both hardware technology and image-processing algorithms. At the component level there have been improvements in scintillators and photon transducers as well as a greater availability of semiconductor technology. These devices permit the fabrication of smaller and more compact systems that can be customized for particular applications. New clinical devices include high-count sensitivity cardiac SPECT systems that do not use conventional collimation and the introduction of diagnostic-quality hybrid SPECT/CT systems. While there has been steady progress with reconstruction algorithms, exciting new processing algorithms have become commercially available that promise to provide substantial reductions in SPECT acquisition time without sacrificing diagnostic quality. Preclinical small-animal SPECT systems have become a major focus in nuclear medicine. These systems have pushed the limits of SPECT into the submillimeter range, making them valuable molecular imaging tools capable of providing information unavailable from other modalities.

Nuclear cystometrogram-determined bladder pressure at onset of vesicoureteral reflux predicts spontaneous resolution.

OBJECTIVES: The bladder pressure at the onset of reflux may reflect the competence of the ureterovesical junction. The predictive value of the bladder pressure at the onset of vesicoureteral reflux, as measured by nuclear cystometrography, was assessed relative to subsequent reflux resolution. METHODS: Nuclear cystometrograms of 67 children were reviewed to determine the presence of reflux, bladder pressure at the onset of reflux, and bladder volume at the onset of reflux. The data were analyzed for significant relationships between the nuclear cystometrogram results and patient outcomes using a Cox proportional hazards regression model, controlling for the duration of follow-up and for patients with bilateral reflux. RESULTS: Data were gathered for reflux occurring in 85 ureters. As expected, reflux of a greater grade was less likely to resolve spontaneously (P = 0.005). The pressure at the onset of reflux was also highly predictive of spontaneous resolution (P = 0.0005). Reflux occurring at greater pressures was more likely to resolve spontaneously, independent of the reflux grade or bladder volume at the onset of reflux. Patient age (P = 0.45), normalized volume of urine refluxed (P = 0.97), and normalized bladder volume at the onset of reflux (P = 0.18) were not significant predictors of reflux resolution. CONCLUSIONS: The findings of this report have demonstrated for the first time, to our knowledge, that the bladder pressure at the onset of vesicoureteral reflux, as determined by nuclear cystometrography, is a significant independent predictor of vesicoureteral reflux resolution in children.

A segmentation-based method for metal artifact reduction.

RATIONALE AND OBJECTIVES: We propose a novel segmentation-based interpolation method to reduce the metal artifacts caused by surgical aneurysm clips. MATERIALS AND METHODS: Our method consists of five steps: coarse image reconstruction, metallic object segmentation, forward-projection, projection interpolation, and final image reconstruction. The major innovations are 2-fold. First, a state-of-the-art mean-shift technique in the computer vision field is used to improve the accuracy of the metallic object segmentation. Second, a feedback strategy is developed in the interpolation step to adjust the interpolated value based on the prior knowledge that the interpolated values should not be larger than the original ones. Physical phantom and real patient datasets are studied to evaluate the efficacy of our method. RESULTS: Compared to the state-of-the-art segmentation-based method designed previously, our method reduces the metal artifacts by 20-40% in terms of the standard deviation and provides more information for the assessment of soft tissues and osseous structures surrounding the surgical clips. CONCLUSION: Mean shift technique and feedback strategy can help to improve the image quality in terms of reducing metal artifacts.