Objective To explore the animal experimental method (AEM) of minimally invasive surgery (MIS) of non-compressive lumbar radiculitis with anti-inflammatory liquid， and evaluate its efficacy with diffusion tensor imaging (DTI) quantitatively.Methods To establish non-compressive lumbar radiculitis model， we injected suspension， coming from autologous nucleus pulpous of 6 Bama miniature pigs.The pigs were divided into two groups （experimental group and control group） randomly into epidural space pararadically at the level of intervertebral foramen of L2~3 and 3~4 as well as L4~5 by CT-guided puncture respectively. 14 days later， we inject anti-inflammatory liquid with same dose into the epidural space in experimental group and normal saline in control group respectively at the same level. Another 3 days and 7 days as well as 14 days later， one piggy in the two groups was sacrificed randomly respectively for TNF-α immunohistochemical study of the nervous root. DTI has been done pre-modeling and prior treatment as well as before sacrifice. Statistic analyses of FA value of the nervous roots had been implemented at the point of execution above， and of TNF-α value post-injection of suspension of anti-inflammation or saline as well.Results With comparison to during pre-modeling period， the FA values of nerve roots are reduced significantly (P<0.05) during post-modeling period in two groups. But， the FA values of nerve roots are elevated gradually in the experimental group on the other hand in the control group are kept non-variable on the 3 days and 7 days as well as 14 days after injection (P>0.05)， and are higher (P<0.05) in the experimental group than in the control group at the same date. On the date of 7 and 14， TNF-α value in the experimental group was decreased (P<0.05).The pigs were and lower than in the control group (P<0.05) on the other hand it kept stable in the control group(P>0.05).Conclusion MIS of non-compressive lumbar radiculitis with anti-inflammatory liquid is effective and， can be evaluated with DTI quantitatively.