Intravertebral tuberculous granuloma is clinically rare and manifests mainly as symptoms of nerve root irritation and spinal cord compression with neurological deficits in addition to symptoms of tuberculosis toxicity. Magnetic resonance imaging is the preferred and most effective imaging method, which does not have a specific segment of occurrence, and MRI shows a proliferative manifestation of intravertebral canal inflammation. Due to the close relationship of adjacent structures, it is often misdiagnosed on imaging as dural proliferative granulomatous inflammation, which is confirmed not to be a dural lesion during surgery. Granulomas are located in the subdural space, which is clearly demarcated from the dura mater and has no obvious adhesions, and there is a relatively clear demarcation between the granuloma and the spinal cord. Therefore, we believe that it is more likely that granulomatous arachnoiditis is caused by the deposition of Mycobacterium tuberculosis in the arachnoid membrane, which would explain the obstruction of cerebrospinal fluid circulation in the subarachnoid space and the markedly increased protein values in cerebrospinal fluid tests. This is because if the granuloma is located at the dural level, it is unlikely to increase the protein value of the cerebrospinal fluid simply because of mechanical compression due to an epidural subdural granuloma, even if it is able to cause impaired cerebrospinal fluid circulation. Under the premise of good blood flow and non-drug resistance, medical treatment should theoretically be effective, but in practice, lesions in the spinal canal gradually form and worsen while lesions in other areas improve during the course of anti-tuberculosis drug therapy, suggesting a more plausible explanation for the downward spread of tuberculosis bacilli through the cerebrospinal fluid and their gradual deposition in the arachnoid, eventually forming granulomas. We consider that the main reason for the ineffectiveness of conservative medical treatment is that the cerebrospinal fluid circulation pathway is blocked by the tuberculosis granuloma, and neither the cerebrospinal fluid nor the blood-transported drugs can reach the lesion area, which explains why the pressure shown by lumbar puncture does not correlate well with the intracranial pressure, and the protein content in the cerebrospinal fluid is significantly higher than the normal value several times. In addition, it should be noted that whether or not the symptoms of tuberculosis toxicity are controlled is not an absolute indication for surgery. In one patient who still had fever with elevated C-reactive protein, the temperature was gradually controlled by continued antituberculosis treatment 1 week after surgery. It indicates that surgical excision and decompression of the lesion and direct apposition with blood-rich muscle tissue can improve the effect of local antituberculosis treatment, and the mechanism may be related to the increase of local blood drug concentration. Regarding the surgical method, domestic Zhang Wende et al. had reported one case of complete resection of an extramedullary subdural granuloma of 12 cm in length from cervical segment 6 to thoracic segment 7, and achieved better results. In our opinion, the total tangible quality on the naked eye is still a partial resection. On MRI, most of the cases showed obvious posterior spinal cord-like hyperplastic lesions in the sagittal position, but the axial scan revealed that the granuloma was not attached to the posterior part of the spinal cord in the form of cords, but was wrapped around the whole spinal cord in a circular barrel. Thus, whatever the extent of resection seen by the naked eye during surgery is only a partial resection of the posterior spinal cord lesion, and with the limited exposure of the laminae removal, a total resection is not possible. Therefore, it is not desirable or possible to perform surgery with the aim of total lesion excision, and it is theoretically unreasonable for the lesion to be located only in the posterior part of the spinal cord, with no lateral anterior infection. In our opinion, the so-called total excision of the lesion is often a misjudgment. Most of the cases in this group have long cumulative segments, and we believe that most of these patients have multi-segmental lesions, and even if the surgery can achieve “total excision” of the posterior spinal cord lesion, there are many problems. Third, the risk of cerebrospinal fluid leakage is easily increased after surgery, which is also the most serious problem, because if all the lesions in the posterior spinal cord within visual range are removed, the cerebrospinal fluid circulation must be opened, and the dura must be sutured to prevent postoperative cerebrospinal fluid leakage. However, closing the dura mater will inevitably lead to the formation of the “blood-spinal cord” barrier again, and postoperative anti-tuberculosis drugs will still be unable to reach the lesion area, which greatly increases the probability of recurrence, and the decompression effect is worse than that of opening the dura mater. Therefore, we believe that the extent of lesion excision should never exceed the upper and lower boundaries of the lesion, and the standard is to not open the cerebrospinal fluid circulation and not to see the outflow of cerebrospinal fluid during the operation. This not only enlarges the vertebral canal cavity to play a bony decompression role, but also provides muscle blood supply to the lesion area and breaks the “blood-spinal cord” barrier, so that anti-tuberculosis drugs can reach the lesion area and thus take effect. In addition, the surface area of the granuloma was significantly increased due to the excision of the granuloma, which provided a larger area for the attachment of the “squid curl” or “checkerboard grid”-like granuloma. This provided a larger area of muscular blood flow to the granuloma, thus allowing anti-tuberculosis drugs to be administered directly to the granuloma lesion itself, avoiding the problem of poor intrathecal anti-tuberculosis drug injection due to poor cerebrospinal fluid circulation. Therefore, although the granuloma itself was not removed, the continued postoperative application of antituberculous drugs was ideal. In addition, we observed from the available cases that the lesions in the non-operative segments located at the upper and lower poles also became smaller and better after surgery, but the results were slightly worse than those in the operative segments, which also supports that the fundamental purpose and key to surgery is to establish effective blood flow to the lesion area without total excision of the lesion, so that the anti-tuberculosis drugs can arrive effectively. In conclusion, although the overall incidence of intravertebral tuberculosis is low, the recent identification of such cases may suggest a trend toward an increased incidence, the cause of which is not yet clear. We speculate that the yearly increase in the incidence of refractory, drug-resistant tuberculous meningitis in recent years may be the underlying cause. Due to the lack of previous knowledge of such patients, the optimal time for surgery is often missed in medical treatment, or even not considered or abandoned, resulting in a reduced degree of neurological recovery and prolonged recovery time. However, total laminectomy has a long-term impact on the stability of the spine, while hemi-laminectomy can significantly reduce the damage to the spinous process, ligaments and small joints, which is beneficial to the stability of the spine and the maintenance of normal physiological function of the spine, but the contact area between the lesion granuloma and the normal tissue is significantly reduced after hemi-laminectomy, and the normal tissue can not be effectively adhered to the surface of the lesion due to the blockage of the spinous process and part of the vertebral plate and the support of the bone thickness. However, the contact area between the lesion granuloma and normal tissue was significantly reduced after hemilaminectomy.