Response of pyogenic spine infection to antibiotic therapy is usually based on nonspecific symptoms and inflammation markers. Abnormalities on MRI persist too long to influence therapy. Is FDG-PET/CT a timely and robust predictor of successful therapy?

Retrospective study. Sequential FDG-PET/CTs done to assess treatment response over a 4-year period. Recurrence of infection after stopping treatment was the endpoint.

One hundred seven patients enrolled. First treatment response scan showed no signs of infection in 69 patients (low risk). Twenty-four additional patients underwent additional treatment after an initial positive scan with low-risk pattern on follow-up imaging. After stopping antibiotics, none had clinical recurrence of infection. One had positive cultures at surgery for negative predictive value of 0.99. Thirty-eight patients had evidence of residual infection. Abnormalities in 28 were comparable to what is seen with untreated infection (high-risk). Twenty-seven received additional treatment until resolution. Antibiotics were stopped in 1 who suffered recurrence.

Ten had low-grade/localized abnormalities consistent with infection (intermediate-risk). Signs of infection resolved in 3 after additional treatment. Of the remaining 7 patients who had minor residual abnormalities when antibiotics were stopped, 1 had recurrent infection for a positive predictive value of 0.14.

Risk stratification proposed: A low-risk scan with only inflammation at a destroyed joint indicates negligible risk of recurrence. Unexplained activity in bone, soft tissue or spinal canal indicates high risk with further antibiotics recommended. Most patients with subtle or localized findings (intermediate risk) did not experience recurrence. Stopping therapy could be considered under careful observation.