Comparison of Commercial AI Software Performance for Radiograph Lung Nodule Detection and Bone Age Prediction.

Background Multiple commercial artificial intelligence (AI) products exist for assessing radiographs; however, comparable performance data for these algorithms are limited. Purpose To perform an independent, stand-alone validation of commercially available AI products for bone age prediction based on hand radiographs and lung nodule detection on chest radiographs. Materials and Methods This retrospective study was carried out as part of Project AIR. Nine of 17 eligible AI products were validated on data from seven Dutch hospitals. For bone age prediction, the root mean square error (RMSE) and Pearson correlation coefficient were computed. The reference standard was set by three to five expert readers. For lung nodule detection, the area under the receiver operating characteristic curve (AUC) was computed. The reference standard was set by a chest radiologist based on CT. Randomized subsets of hand (n = 95) and chest (n = 140) radiographs were read by 14 and 17 human readers, respectively, with varying experience. Results Two bone age prediction algorithms were tested on hand radiographs (from January 2017 to January 2022) in 326 patients (mean age, 10 years ± 4 [SD]; 173 female patients) and correlated strongly with the reference standard (r = 0.99; P < .001 for both). No difference in RMSE was observed between algorithms (0.63 years [95% CI: 0.58, 0.69] and 0.57 years [95% CI: 0.52, 0.61]) and readers (0.68 years [95% CI: 0.64, 0.73]). Seven lung nodule detection algorithms were validated on chest radiographs (from January 2012 to May 2022) in 386 patients (mean age, 64 years ± 11; 223 male patients). Compared with readers (mean AUC, 0.81 [95% CI: 0.77, 0.85]), four algorithms performed better (AUC range, 0.86-0.93; P value range, <.001 to .04). Conclusions Compared with human readers, four AI algorithms for detecting lung nodules on chest radiographs showed improved performance, whereas the remaining algorithms tested showed no evidence of a difference in performance. © RSNA, 2024 Supplemental material is available for this article. See also the editorial by Omoumi and Richiardi in this issue.

[Which imaging methods are safe in pregnant patients? A step-by-step plan to choose the right method]. / Welk beeldvormend onderzoek bij een zwangere patiënt?

Diagnostic imaging in pregnant patients often causes worries about radiation effects and IV contrast. We use a step-by-step plan to choose the right method and illustrate this with a case. First, we choose the method of imaging, non-ionizing imaging being preferred over ionizing. With ionizing radiation, the fetal dose is limited as much as possible with teratogenic and carcinogenic risk as low as possible (as low as reasonably achievable (ALARA)). The second consideration is whether intravenous contrast medium is necessary. The risks of the imaging procedure for the fetus (and patient) are compared with the risks for the patient (and fetus) without imaging. Furthermore, possible treatment changes are important. Finally, the patient is informed about the procedure, the necessity of it and possible negative effects for the fetus and patient. In our case these risks were minimal and it was not necessary to explicitly discuss them with the patient.

Predicting no return to sports after three months in patients with traumatic knee complaints in general practice by combining patient characteristics, trauma characteristics and knee complaints.

Background: It remains unclear to what extent patients with traumatic knee complaints aged 18-45 years seen in general practice experience difficulties with return to sports.Objectives: This study aims to determine the proportion of patients with a knee trauma that return to sports at six weeks and three months follow-up. Also examined were associations between no return to sports and baseline patient/trauma characteristics, knee complaints and MR (magnetic resonance) findings, as well as the additive value of MR findings.Methods: Included were patients with traumatic knee complaints participating in a randomized controlled trial assessing the cost-effectiveness of an MR scan in general practice. Patients were classified as 'no return to sports' or 'return to sports' (sports on pre-injury or adapted level). Potential baseline predictors for no return to sports were assessed using logistic regression analyses. The area under the curves (AUC) was compared.Results: At six weeks and three months follow-up, 147 (59%) and 175 (74%) patients, respectively, reported return to sports. Combining patient characteristics, trauma characteristics and knee complaints predicted no return to sports with an AUC of 0.86 (95%CI: 0.81-0.90) at six weeks and of 0.82 (95%CI: 0.76-0.88) at three months follow-up. After adding MR findings, the AUC was 0.79 (95%CI: 0.71-0.87) at six weeks and 0.79 (95%CI: 0.70-0.88) at three months follow-up.Conclusion: Three out of four patients with a knee trauma in general practice reported return to sports at three months follow-up. A combination of patient/trauma characteristics and knee complaints predicted no return to sports, whereas MR findings had no additive value. Trial registration: Dutch trial registration: registration number: NTR3689. registration date: 7 November 2012.

Does MRI add value in general practice for patients with traumatic knee complaints? A 1-year randomised controlled trial.

OBJECTIVE: To determine whether referral to MRI by the general practitioner (GP) is non-inferior to usual care (no access to MRI by GPs) in patients with traumatic knee complaints regarding knee-related daily function. METHODS: This was a multicentre, non-inferiority randomised controlled trial with 1-year follow-up. GPs invited eligible patients during or after their consultation. Eligible patients (18-45 years) consulted a GP with knee complaints due to a trauma during the previous 6 months. Patients allocated to the MRI group received an MRI at (median) 7 (IQR 1-33) days after the baseline questionnaire. Patients in the usual care group received information on the course of knee complaints, and a referral to a physiotherapist or orthopaedic surgeon when indicated. The primary outcome measure was knee-related daily function measured with the Lysholm scale (0 to 100; 100=excellent function) over 1 year, with a non-inferiority margin of 6 points. RESULTS: A total of 356 patients were included and randomised to MRI (n=179) or usual care (n=177) from November 2012 to December 2015. MRI was non-inferior to usual care concerning knee-related daily function during 1-year follow-up, for the intention-to-treat (overall adjusted estimate: 0.33; 95% CI -1.73 to 2.39) and per-protocol (overall adjusted estimate: 0.06; 95% CI -2.08 to 2.19) analysis. There were no differences between both groups in the amount of patients visiting other healthcare providers. CONCLUSION: MRI in general practice in patients with traumatic knee complaints was non-inferior to usual care regarding knee-related daily function during 1-year follow-up. TRIAL REGISTRATION NUMBER: NTR3689.

Clinical management of spinal metastases-The Dutch national guideline.

This article is a summary of the revised Dutch multidisciplinary evidence-based guideline 'Spinal metastases' (English translation available at: https://www.oncoline.nl/spinal-metastases) that was published at the end of 2015. This summary provides an easy-to-use overview for physicians to use in their daily practice.

The Dutch national guideline on metastases and hematological malignancies localized within the spine; a multidisciplinary collaboration towards timely and proactive management.

Here, we describe the development of a Dutch national guideline on metastases and hematological malignancies localized within the spine. The aim was to create a comprehensive guideline focusing on proactive management of these diseases, enabling healthcare professionals to weigh patient perspectives, life expectancy, and expected outcomes to make informed treatment recommendations. A national multidisciplinary panel consisting of clinicians, a nurse, a patient advocate, an epidemiologist, and a methodologist drafted the guideline. The important role of patients in the realization of the guideline enabled us to identify and address perceived shortcomings in patient care. The guideline covers not only metastatic epidural spinal cord compression, but also the treatment of uncomplicated metastases and hematological malignancies localized within the spine. The guideline is applicable in daily practice and provides an up-to-date and concise overview of the diagnostic and treatment possibilities for patients suffering from a disease that can have a serious impact on their quality of life. Suggestions for the practical implementation of patient care in hospitals are also provided, including approaches for pursuing proactive management. The crucial role of the patient in decision making is emphasized in this guideline.

Combining independent decisions increases diagnostic accuracy of reading lumbosacral radiographs and magnetic resonance imaging.

Diagnosing the causes of low back pain is a challenging task, prone to errors. A novel approach to increase diagnostic accuracy in medical decision making is collective intelligence, which refers to the ability of groups to outperform individual decision makers in solving problems. We investigated whether combining the independent ratings of chiropractors, chiropractic radiologists and medical radiologists can improve diagnostic accuracy when interpreting diagnostic images of the lumbosacral spine. Evaluations were obtained from two previously published studies: study 1 consisted of 13 raters independently rating 300 lumbosacral radiographs; study 2 consisted of 14 raters independently rating 100 lumbosacral magnetic resonance images. In both studies, raters evaluated the presence of "abnormalities", which are indicators of a serious health risk and warrant immediate further examination. We combined independent decisions of raters using a majority rule which takes as final diagnosis the decision of the majority of the group. We compared the performance of the majority rule to the performance of single raters. Our results show that with increasing group size (i.e., increasing the number of independent decisions) both sensitivity and specificity increased in both data-sets, with groups consistently outperforming single raters. These results were found for radiographs and MR image reading alike. Our findings suggest that combining independent ratings can improve the accuracy of lumbosacral diagnostic image reading.

General Practitioners Referring Adults to MR Imaging for Knee Pain: A Randomized Controlled Trial to Assess Cost-effectiveness.

Purpose To determine the cost-effectiveness of early referral by the general practitioner for magnetic resonance (MR) imaging compared with usual care alone in patients aged 18-45 years with traumatic knee symptoms. Materials and Methods Cost-utility analysis was performed parallel to a prospective multicenter randomized controlled trial in Dutch general practice. A total of 356 patients with traumatic knee symptoms were included from November 2012 to December 2015 (mean age, 33 years ± 8 [standard deviation]; 222 men [62%]). Patients were randomly assigned to usual care (n = 177; MR imaging was not performed, but patients were referred to an orthopedic surgeon when conservative treatment was unsatisfactory) or MR imaging (n = 179) within 2 weeks after injury. Main outcome measures were quality-adjusted life years (QALYs) and costs from a healthcare and societal perspective. Multiple imputation was used for missing data. The Student t test was used to assess differences in mean QALYs, costs, and net benefits. Results Mean QALYs were 0.888 in the MR imaging group and 0.899 in the usual care group (P = .255). Healthcare costs per patient were higher in the MR imaging group (€1109) than in the usual care group (€837) (P = .050), mainly due to higher costs for MR imaging, with no reduction in the number of referrals to an orthopedic surgeon in the MR imaging group. Conclusion MR imaging referral by the general practitioner was not cost-effective in patients with traumatic knee symptoms; in fact, MR imaging led to more healthcare costs, without an improvement in health outcomes.

Post-traumatic knee MRI findings and associations with patient, trauma, and clinical characteristics: a subgroup analysis in primary care in the Netherlands.

BACKGROUND: The added value of magnetic resonance imaging (MRI) in primary care is still being debated. A high diagnostic yield can be expected in young and active patients with post-traumatic knee complaints. AIM: To determine the frequency of MRI abnormalities in young and active patients (aged 18-45 years) and the associations with patient, trauma, and clinical characteristics. DESIGN AND SETTING: A subgroup analysis of 174 patients, aged 18-45 years with knee trauma of <6 months, allocated to MRI in a randomised controlled trial on the yield of MRI in primary care. Patients were recruited by 150 GPs in the Netherlands from October 2012 to November 2015. METHOD: Associations were expressed using mean differences, odds ratio (OR) and predictive values. RESULTS: Sixty-seven out of 174 patients (39%) had a positive MRI finding, predominantly anterior cruciate ligament (ACL) ruptures (22%) and/or traumatic meniscal tears (15%). Patients with a pre-existing musculoskeletal comorbidity had a two-fold lower prevalence of positive MRI findings (21%), OR 3.0 (95% confidence interval [CI] = 1.3 to 7.0). A 'sports related trauma' showed the highest OR of 4.6 (95% CI = 2.2 to 9.3) for a positive MRI finding. Clinical scores were statistically, significantly worse in patients with positive MRI findings, with mean differences ranging from 10 to 20%. Furthermore, increasing duration of complaints was correlated with decreasing prevalence rates of positive MRI findings. Overall, a popping sound and direct swelling showed the highest positive predictive value of 65% for the presence of positive MRI findings. CONCLUSION: The results from this study enable a preselection of patients to increase the diagnostic yield of MRI in primary care.

Direct access to magnetic resonance imaging improved orthopaedic knee referrals in the Netherlands.

BACKGROUND: The use of magnetic resonance imaging (MRI) in primary care is under debate, and the majority of GPs have no experience with MRI. OBJECTIVES: To examine for which patients with knee injury an MRI is ordered and does direct access to MRI in primary care influence the GP referral to an orthopaedic surgeon? METHODS: Consecutive patients with knee injury who underwent an MRI examination ordered by their GP were included. On the application form for MRI, the GPs indicated their referral intention in advance, as if MRI had not been available. Six months after the MRI scan, written interviews with the GPs were used to collect data on referrals and orthopaedic intervention. The number of patients finally referred to an orthopaedic surgeon in secondary care after MRI was compared with the number of intended referrals. RESULTS: Of the 588 included, GPs referred fewer patients to the orthopaedic surgeon after receiving the MRI results than they would have done prior to MRI (60% versus 82.8%, P < 0.0001). The reduction was 16.1% for patients older than 50 years and 28.1% for patients younger than 50 years. Orthopaedic intervention was performed in 62.9% of all referred patients. Of the 101 patients whom the GP did not intend to refer prior to MRI, 48 were referred to an orthopaedic surgeon based on the MRI findings. CONCLUSION: In patients with knee injury, direct access to MRI of the knee in a primary care setting significantly reduced referrals to an orthopaedic surgeon. LEVEL OF EVIDENCE: Three prospective cohort.

Improving your approach to nasal obstruction.

The causes are diverse--from rhinitis and rhinosinusitis to drugs and structural/mechanical abnormalities. Here's how to provide patients with relief.

Diagnostic Accuracy of Lumbosacral Spine Magnetic Resonance Image Reading by Chiropractors, Chiropractic Radiologists, and Medical Radiologists.

STUDY DESIGN: A cross-sectional diagnostic accuracy study was conducted in 2 sessions. OBJECTIVE: It is important to know whether it is possible to accurately detect "specific findings" on lumbosacral magnetic resonance (MR) images and whether the results of different observers are comparable. SUMMARY OF BACKGROUND DATA: Health care providers frequently use magnetic resonance imaging in the diagnostic process of patients with low back pain. The use of MR scans is increasing. This leads to an increase in costs and to an increase in risk of inaccurately labeling patients with an anatomical diagnosis that might not be the actual cause of symptoms. METHODS: A set of 300 blinded MR images was read by medical radiologists, chiropractors, and chiropractic radiologists in 2 sessions. Each assessor read 100 scans in round 1 and 50 scans in round 2. The reference test was an expert panel.For all analyses, the magnetic resonance imaging findings were dichotomized into "specific findings" or "no specific findings." For the agreement, percentage agreement and κ values were calculated and for validity, sensitivity, and specificity. Sensitivity analysis was done for classifications A and B (prevalence of 31% and 57%, respectively). RESULTS: The intraobserver κ values for chiropractors, chiropractic radiologists, and medical radiologists were 0.46, 0.49, and 0.69 for A and 0.55, 0.75, and 0.64 for B, respectively.The interobserver κ values were lowest for chiropractors (0.28 for A, 0.37 for B) and highest for chiropractic radiologists (0.50 for A, 0.49 for B).The sensitivities of the medical radiologists, chiropractors, and chiropractic radiologists were 0.62, 0.71, and 0.75 for A and 0.70, 0.74, 0.84 for B, respectively.The specificities of medical radiologists, chiropractic radiologists, and chiropractors were 0.82, 0.77, and 0.70 for A and 0.74, 0.52, and 0.61 for B, respectively. CONCLUSION: Agreement and validity of MR image readings of chiropractors and chiropractic and medical radiologists is modest at best. This study supports recommendations in clinical guidelines against routine use of magnetic resonance imaging in patients with low back pain. LEVEL OF EVIDENCE: 3.

Efficacy of MRI in primary care for patients with knee complaints due to trauma: protocol of a randomised controlled non-inferiority trial (TACKLE trial).

BACKGROUND: Patients with traumatic knee complaints regularly consult their general practitioner (GP). MRI might be a valuable diagnostic tool to assist GPs in making appropriate treatment decisions and reducing costs. Therefore, this study will assess the cost-effectiveness of referral to MRI by GPs compared with usual care, in patients with persistent traumatic knee complaints. DESIGN AND METHODS: This is a multi-centre, open-labelled randomised controlled non-inferiority trial in combination with a concurrent observational cohort study. Eligible patients (aged 18-45 years) have knee complaints due to trauma (or sudden onset) occurring in the preceding 6 months and consulting their GP. Participants are randomised to: 1) an MRI group, i.e. GP referral to MRI, or 2) a usual care group, i.e. no MRI. Primary outcomes are knee-related daily function, medical costs (healthcare use and productivity loss), and quality of life. Secondary outcomes are disability due to knee complaints, severity of knee pain, and patients' perceived recovery and satisfaction. Outcomes are measured at baseline and at 1.5, 3, 6, 9 and 12 months follow-up. Also collected are data on patient demographics, GPs' initial working diagnosis, GPs' preferred management at baseline, and MRI findings. DISCUSSION: In the Netherlands, the additional diagnostic value and cost-effectiveness of direct access to knee MRI for patients presenting with traumatic knee complaints in general practice is unknown. Although GPs increasingly refer patients to MRI, the Dutch clinical guideline 'Traumatic knee complaints' for GPs does not recommend referral to MRI, mainly because the cost-effectiveness is still unknown. TRIAL REGISTRATION: Dutch Trial Registration: NTR3689.

[Using functional MRI to measure happiness]. / Geluk meten met functionele MRI.

In addition to MRI for use in diagnostic imaging we now also have functional MRI (fMRI) at our disposal. This can be used to map activity in different areas of the brain. Since its introduction, this technique has been used extensively in preoperative analysis of brain tumours. Emotional centres can also be mapped. Various applications and limitations of fMRI for studying emotions are discussed in this article.

[MRI diagnostics should be available to general practice]. / MRI-diagnostiek hoort ook bij de huisarts.

With its high negative predictive value, MRI is of great value to the general practitioner's (GP) unequivocal task of excluding pathology in a large proportion of their patient population. While GPs prefer open access to MRI, insurance companies prefer to limit or discourage GP use of MRI, the main fear being unlimited unnecessary requests. Studies have shown these fears to be unfounded. Dutch Health Care Insurance Board (CVZ) regulations stipulate that GPs should have access to MRI on specific indications that are described in the Dutch College of General Practitioners' (NHG) practice guidelines. In the Netherlands, Alkmaar Medical Center (MCA) has offered open access to MRI since 1993 (knee) and 2006 (hernia nuclei pulposi), with excellent results. The percentage of normal report rates and of pathological findings of GPs is comparable to that of both neurologists and orthopaedic surgeons. Open access to MRI helps to speed up the diagnostic process, reduce outpatient waiting lists and limit unnecessary consultations. International publications have reported the cost-effectiveness of MRI. Experience both outside the Netherlands and at MCA has shown that open access to MRI results in reduced waiting lists and enhanced cost-effectiveness. This is best practice and should be incorporated into NHG practice guidelines.

Assessment of microvascular obstruction and prediction of short-term remodeling after acute myocardial infarction: cardiac MR imaging study.

PURPOSE: To evaluate which cardiac magnetic resonance (MR) imaging technique for detection of microvascular obstruction (MVO) best predicts left ventricular (LV) remodeling after acute myocardial infarction (MI). MATERIALS AND METHODS: This study had local ethics committee approval; all patients gave written informed consent. Sixty-three patients with first acute MI, treated with primary stent placement and optimal medical therapy, underwent cine MR imaging at 4-7 days and at 4 months after MI. Presence of MVO was qualitatively evaluated at baseline by using three techniques: (a) a single-shot saturation-recovery gradient-echo first-pass perfusion sequence (early hypoenhancement), mean time, 1.09 minutes +/- 0.07 (standard deviation) after contrast material administration; (b) a three-dimensional segmented saturation-recovery gradient-echo sequence (intermediate hypoenhancement), mean time, 2.17 minutes +/- 0.26; and (c) a two-dimensional segmented inversion-recovery gradient-echo late gadolinium enhancement sequence (late hypoenhancement), mean time, 13.32 minutes +/- 1.26. Contrast-to-noise ratios (CNRs) were calculated from the signal-to-noise ratios of the infarcted myocardium and MVO areas. Univariable linear regression analysis was used to identify the predictive value of each MR imaging technique. RESULTS: Early hypoenhancement was detected in 44 (70%) of 63 patients; intermediate hypoenhancement, in 39 (62%); and late hypoenhancement, in 37 (59%). Late hypoenhancement was the strongest predictor of change in LV end-diastolic and end-systolic volumes over time (beta = 14.3, r = 0.40, P = .001 and beta = 11.3, r = 0.44, P < .001, respectively), whereas intermediate and late hypoenhancement had comparable predictive values of change in LV ejection fraction (beta = -3.1, r = -0.29, P = .02 and beta = -2.8, r = -0.27, P = .04, respectively). CNR corrected for spatial resolution was significantly superior for late enhancement compared with the other sequences (P < .001). CONCLUSION: By using cardiac MR imaging, late hypoenhancement is the best prognostic marker of LV remodeling, with highest CNR between the infarcted myocardium and MVO regions.

Functional recovery after acute myocardial infarction: comparison between angiography, electrocardiography, and cardiovascular magnetic resonance measures of microvascular injury.

OBJECTIVES: We examined the relation between angiographic, electrocardiographic, and gadolinium-enhanced cardiovascular magnetic resonance (CMR) characteristics of microvascular obstruction (MVO), and their predictive value on functional recovery after acute myocardial infarction (AMI). BACKGROUND: Microvascular obstruction on CMR has been shown to predict left ventricular (LV) remodeling, but it is not well known how it compares with commonly used criteria of microvascular injury, and earlier reports have produced conflicting results on the significance and extent of MVO. METHODS: Thrombolysis In Myocardial Infarction (TIMI) flow grade, myocardial blush grade (MBG), and ST-segment resolution were assessed in 60 patients with AMI treated with primary stenting. Cardiovascular magnetic resonance was performed between 2 and 9 days after revascularization to determine early MVO on first-pass perfusion imaging, late MVO on late gadolinium-enhanced imaging, and infarct size and transmural extent. Cine imaging was used to determine LV volumes and global and regional function at baseline and 4-month follow-up. RESULTS: Early and late MVO were both related to incomplete ST-segment resolution (p = 0.002 and p = 0.01, respectively), but not to TIMI flow grade and MBG. Of all angiographic, electrocardiographic, and CMR variables, late MVO was the strongest parameter to predict changes in end-diastolic volume (beta = 0.53; p = 0.001), end-systolic volume (beta = 8.67; p = 0.001), and ejection fraction (beta = 3.94; p = 0.006) at follow-up. Regional analysis showed that late MVO had incremental diagnostic value to transmural extent of infarction (odds ratio: 0.18; p < 0.0001). CONCLUSIONS: In patients after revascularized AMI, late MVO proved a more powerful predictor of global and regional functional recovery than all of the other characteristics, including transmural extent of infarction.

Late gadolinium-enhanced cardiovascular magnetic resonance evaluation of infarct size and microvascular obstruction in optimally treated patients after acute myocardial infarction.

PURPOSE: Cardiovascular magnetic resonance (CMR) is considered the standard imaging modality in clinical trials to monitor patients after acute myocardial infarction (AMI). However, limited data is available with respect to infarct size, presence, and extent of microvascular injury (MVO), and changes over time, in relation to cardiac function in these optimally treated patients. In this study, we prospectively investigate the change of infarct size over time, and the incidence and significance of MVO in a uniform, optimally treated patient group after AMI. METHODS: Forty patients underwent cine and late gadolinium-enhanced CMR within 9 days and at 4 months after primary stenting. Left ventricular ejection fraction (LVEF), infarct size (IS) and MVO size were calculated. RESULTS: IS decreased with 19.0% at follow-up (p<0.01). The 23 (57.5%) patients with MVO had larger infarct size, higher left ventricular volumes and lower LVEF and more involution of IS at follow-up. Overall, LVEF improved from 42.3+/-9.8% to 44.0+/-9.8% (p=0.06), irrespective of presence or size of MVO. CONCLUSION: Infarct size reduces over time by 19.0% in optimally treated patients after AMI. Despite optimal reperfusion, MVO was found in the majority of patients. Although patients with MVO had larger infarcts and worse indices of left ventricular remodelling, functional change at follow-up was comparable to patients without MVO.

Role of radiology in a national initiative to interdict drug smuggling: the Dutch experience.

OBJECTIVE: The purpose of this pictorial essay is to describe the role of radiology in a national initiative to intercept illegal narcotics concealed within the bodies of human transporters. CONCLUSION: Radiologic examination is increasingly important in identifying intracorporeal drug smuggling as improved wrapping techniques undermine the usefulness of blood and urine testing and clinical observation. Detection rates of high accuracy, sensitivity, and specificity are achieved by experienced radiologists.

Interobserver agreement on the radiological criteria of the International Panel on the diagnosis of multiple sclerosis.

We assessed the interobserver agreement on the radiological part of the International Panel (IP) criteria for the diagnosis of multiple sclerosis (MS), comprising the assessment of dissemination in space (DIS) and time (DIT) based exclusively on MRI. Four radiologists trained and four radiologists naive in the application of the IP criteria scored the fulfillment for DIS (i.e., > or =3 periventricular, > or =1 infratentorial, > or =1 juxtacortical, > or = 1 enhancing lesion or a total of > or =9 T2-weighted lesions) and DIT (presence of new or enhancing lesions at follow-up) in baseline and two follow-up scans from 20 patients suspected for having MS. The IP-trained radiologists agreed at least moderately on all assessments (kappa>0.40), whereas the IP-naive radiologists showed fair agreement (kappa<0.40) on five of 16 assessments. In the final conclusion on DIS and DIT, the IP-trained radiologists agreed substantially on both items (kappa=0.62 and kappa=0.60, respectively) compared with a fair agreement on DIS (kappa=0.29) and moderate agreement on DIT (kappa=0.52) among the IP-naive radiologists. Given the poor interobserver agreement among IP-naive observers, the new IP criteria for MS require additional training and should perhaps be simplified.