Magnetic resonance imaging (MRI) is considered the gold standard diagnostic modality for spinal pathologies [1]. However, relying solely on reports to diagnose spinal conditions is insufficient, as radiologist performance is often imperfect, and some errors are unavoidable [2]. Interobserver variability exceeding 50% has been reported in MRI evaluations of patients with lumbosacral radicular pain [3]. Thus, the treating spine surgeon should meticulously review MRI films, correlate them with the patient’s symptoms, and make an accurate diagnosis. Presently, MRI results are often provided exclusively on compact discs (CDs), which are more suitable for data transmission than for routine clinical analysis. In our region, including the city and country where we practice, MRI centers typically provide physical MRI films (plates). Digital media such as CDs are infrequently used in our region owing to practical limitations such as compatibility issues, the absence of CD drives in modern computers, and the additional time—often twice as long as MRI films—required for viewing, which reduces surgeon productivity in high-demand outpatient settings [4]. Although digital systems such as the Picture Archiving and Communication System are on the rise globally, difficulties, including costly hardware, reliable information technology infrastructure, and the financial burden of maintaining these software systems, are exhausting for budget-tight clinics. Physical MRI films remain a practical and widely used choice in resource-limited settings, particularly in smaller clinics. Therefore, in a clinical setting, physical MRI films are standard practice. In such a situation, variability in the quality of available images can be readily observed, with numerous inconsistencies and limitations frequently impacting precise decision-making and patient outcomes. As the results of MRI examinations are critical for payers, determining the acceptability of prescribed treatments, an accurate diagnosis is crucial for prompt and effective intervention. For these reasons, we undertook this study to examine the deficiencies in MRI films provided by diagnostic centers, which may lead to questionable and inconclusive diagnoses.

The study was conducted after approval from the institutional review board of Bombay Hospital and Medical Research Centre, Mumbai (BH-EC-0127) and obtaining informed consent from the patients. MRI film sets of the lumbosacral spine brought by patients from more than 100 MRI centers between January 2023 and March 2024 were included in this study and assessed for technical inadequacies. Only film sets with comprehensive imaging, excluding screening films, were considered. The evaluation parameters included the number of films provided, T1 axial and T1 sagittal images, side-specific sagittal image sequence profiling (noting right-to-left [R-L] or left-to-right [L-R] designations), adherence to Macnab’s rule for sectioning, parallel axial sectioning of scout films relative to the examined segment, and sacroiliac (SI) joint screening.

A spreadsheet was created to record these parameters, which was updated following the inspection of each MRI film set in the outpatient department. The results were compiled, detailing the percentage and number of MRI sets lacking the studied parameters.

A total of 1,150 lumbar MRI sets from more than 100 MRI centers were examined for technical inadequacies. The findings showed that 35% of the sets did not include T1 axial images, whereas 8% lacked T1 sagittal images. In 38% of the sets, sagittal image sequencing was not specified as being from right to left or from left to right. Of those that provided sequencing information, 85% used a R-L profile, and 15% used a L-R profile. Macnab’s recommendations for sectional levels at each segment were not adhered to in 970 sets. Additionally, 350 sets had nonparallel axial sectioning of the scout films relative to the examined segment. In addition, 40% of the MRI sets did not include SI joint screening. The number of films provided by the MRI centers ranged from two to six.

The level of radiologists’ specialization, the types of equipment used the lack of involvement in film issuance, and the variability in terminology for MRI findings all contribute to differences in interpretation and the prevalence of errors. This study primarily addressed the practical challenges faced by spine surgeons in analyzing MRI films, developing treatment plans, and offering treatment options to patients. We also considered the difficulties encountered during the execution of these treatment plans because of the inadequacies of MRI film sets. Below are practical examples of each evaluated parameter and their impact on clinical practice.

T1 axial images: 35% of the MRI film sets did not have T1 axial images.

Example 1: A patient presented with chronic right-sided leg symptoms. However, T2 axial imaging revealed bilateral hypointensity in the spinal canal, raising unnecessary concerns about potential pathology on the left side despite the patient’s symptoms being confined to the right side (Fig. 1). Although our intention was to approach and treat the right side alone, this T2 axial image created unnecessary anxiety in the form of compression on the asymptomatic side. Fortunately, T1 axial imaging provided clarity, showing that the pathology was limited to the right side and aligned with the patient’s clinical presentation. This was reassuring and conclusive, confirming that the left side was unaffected and enabling a targeted symptom-based surgical plan with confidence. Literature and experience have demonstrated that it is not unusual to see a patient with ipsilateral disc herniation and contralateral symptoms [5].

Example 2: An elderly male patient with severe claudication pain presented with the following images (T2 axial and sagittal images), with hyperintensity in both images (Fig. 2A). The T1 axial image showed stenosis with a hyperintense area in the canal (Fig. 2B). This helped us make a spot diagnosis of epidural lipomatosis. Without the T1 axial image, both diagnosis and treatment would have been incorrect.

T1 Sagittal images: 8% of the MRI film sets did not include T1 sagittal images.

Example: A T2 axial image shows significant stenosis (Fig. 3A). What would be the treatment plan? In such cases, there is often uncertainty about whether to perform decompression alone or to include fusion if T1 sagittal images are unavailable. T1 sagittal images are crucial as they reveal key features, such as the vacuum sign and Modic changes, which help determine the appropriate treatment plan. A prominent vacuum sign on the T1 sagittal image indicated that fusion would be preferred over decompression (Fig. 3B).

Modic changes cannot be accurately assessed without T1 sagittal images. There are three types of Modic changes. Type I appears as hypointensity on T1 images and hyperintensity on T2 images (Fig. 4A), type II shows hypointensity on both T1 and T2 images (Fig. 4B), and type III shows hyperintensity on both T1 and T2 images (Fig. 4C).

In our study, 350 MRI sets were sectioned nonparallel to the endplates, which is a significant issue. Nonparallel sections can lead to misleading information about both the anatomy and pathology at the affected level.

Example: A patient with left-leg symptoms and spondylolytic spondylolisthesis had an intraoperative finding in which the nerve root was precisely positioned over the disc. This could be clearly identified during preoperative planning owing to the parallel sectioning performed. The MRI image shows that the nerve root had been displaced and positioned in front of the disc (Fig. 5A). Without parallel sectioning, preoperative planning is challenging and could lead to disastrous intraoperative events, including potential nerve root damage. Fig. 5B and C demonstrate examples of parallel versus nonparallel sectioning.

The Macnab rule was not followed in 970 sets.

According to Macnab’s rule, MRI sectioning should be performed at four key levels: suprapedicular, pedicular, infrapedicular/foraminal, and discal (Fig. 6A). Failing to adhere to this rule is problematic because it results in missing critical information needed for accurate MRI analysis and decision-making. Pathologies at the suprapedicular level, such as disc migration from the upper level or compression of the traversing nerve root, may go unnoticed. Understanding pedicle size, orientation, and anatomy requires sectioning at the pedicle level, whereas foraminal pathologies necessitate infrapedicular sections. Fig. 6B shows where Macnab’s rule was not followed.

Sagittal image sequencing in MRI films is crucial for accurately identifying the side related to the pathology. Our study found that 38% of the MRI films did not specify the direction of the sagittal image sequencing (R-L/L-R) (Fig. 7A). Among those that did, 85% were oriented from R-L, while 15% were from L-R (Fig. 7B). This lack of standardization resulted in inconsistent sequencing. Our study revealed a significant discrepancy in the number of MRI films provided, with MRI centers supplying between two and six films. For instance, in Fig. 8, only two films were provided, with a disorganized presentation of the images. This haphazard arrangement makes it challenging for surgeons to analyze incomplete images, which are neither visually clear for identifying pathologies nor mentally satisfactory for interpretation.

The SI joint serves as a crucial connection between the spine and pelvis, making its screening essential. In 2020, approximately 619 million people globally experienced low back pain, and this number is expected to rise to 843 million by 2050, where in 10%–25% of low back pain cases, the SI joint is the source of pain [6]. SI joint pathologies often mimic spinal issues, such as buttock, thigh, and S1 dermatomal pain [7]. Our study revealed that 40% of the MRI sets did not include SI joint screening, which is a significant oversight. Without proper SI joint screening, such pathologies are frequently overlooked or underdiagnosed, leading to inadequate treatment.

In addition to the aforementioned limitations of the MRI sets, our study revealed that many images were exceptionally small. This often necessitated the use of magnifying glasses to properly analyze and determine the diagnosis.

Based on the results obtained, we strongly recommend that radiologists establish a consensus to achieve uniformity, address inadequacies, maximize precision, and minimize the chances of error, which should lead to improved diagnosis and clinical outcomes.