We are pleased that our article “Diffusion-Weighted Imaging in the Assessment of Brain Abscesses Therapy” has resulted in several letters to the editor, and here we will attempt to address the issues raised in those letters.
We agree with Dr. Khramov in that to prove the value of diffusion-weighted (DW) imaging in management of cerebral abscesses would require a more rigorous investigation than ours. Our study was comprised of our “limited” experience in seven patients, a fact that should lead no reader to believe that our investigation is a definite one but, rather, an enthusiastic beginning for what is a complex clinical issue. We hope that our findings will lead other investigators to pursue this exciting issue with more depth. We thank Dr. Khramov for pointing out what can be explained by a typographical error in Table 2. The patient identified as 2 on part B of the table should actually be patient 3. We sincerely apologize for overlooking this mistake. In addition, Dr. Khramov is correct in pointing out other errors in Table 2, particularly where the columns appear to have “shifted” (this resulted in confusion regarding the results listed in the columns labeled “time” and “DWI”). To cast doubt on our results and integrity on the basis of typographical mistakes, however, is unwarranted.
Dr. Khramov goes on to question the relationship between apparent diffusion coefficient (ADC) and DW imaging values. As we stated in our article, the mean ADC in untreated abscesses (representing the initial MR study) was 0.52 (range, 0.36–0.75 × 10−3 mm2/s), and all abscesses had high signal intensity on DW imaging, probably because of restricted diffusion. These findings in our small series correlate with previous experience, our own experience, and that of others. ADC values measured after surgical drainage (first follow-up) showed considerable variations (0.51–2.95 × 10−3 mm2/s), which we believe reflect the contents of the drained cavities (a combination of artificial CSF, antibiotics, fluids, and hemorrhage). The heterogeneity of the cavities may have accounted for the overlap in the values we obtained.
Furthermore, Dr. Khramov is correct in pointing out that we did not state whether DW imaging, ADC, or both was the deciding value for performing drainage. We did not seek to establish a threshold ADC but cautiously concluded that signal intensity changes on DW imaging (in the correct clinical and laboratory settings) may detect the reappearance of pus before conventional MR imaging. The lack of clinical deterioration and decreasing C-reactive protein in patient 7 explains surgery was not performed again. In patient 2, it is obvious that, despite relatively decreasing ADC values (less restriction) in the abscess cavity, his C-reactive protein continued to rise, he was showing a lack of clinical improvement, and the DW image showed very high signal intensity, all factors that lead to repeat surgery. Therefore, we are well aware of the complexity in the management of these patients and are not advocating the use of DW imaging/ADC independent of the other clinical parameters. We do not agree with Dr. Khramov’s claim that these tendencies contradict our hypothesis. To say that our “early” experience is simply narrative and that our study is flawed represents, in our opinion, a lack of understanding of the complexity involved in managing these type of patients.
The second letter takes a more benign tone. Dr. Duprez tells us about a patient who is interesting because a partially treated abscess showed increasing ADC values despite aspiration that confirmed the presence of pus. It is impossible to know what changes in the microenvironment of the pus are introduced by antibiotics (it is not even clear whether antibiotics truly penetrate through the abscess capsule into the pus) or to what extent T2 relaxation effects may account for these findings. Antibiotics must have an effect on the pus, because in many partially treated abscesses the responsible microorganism is never isolated from aspirated pus. In his second patient, high ADC was also found in the pus of another partially treated abscess. The role (if any) that bacteria play in the signal intensity characteristics of abscesses remains to be elucidated. In addition, these two cases also show the complex interplay between signal intensity characteristics on DW imaging and T2 relaxation effects. Dr. Duprez also calls our attention to the different signal intensities between the peripheral and central aspects of the lesions. Although we did not mention this observation in our article, just by glancing at our illustrations one can observe a hypo-/hyperinterface (such as the one he mentions) on the DW image in untreated abscesses and in the recurrent abscess shown in Figure 4. This interface, however, is absent in the successfully treated abscesses. We thank Dr. Duprez for pointing out this finding and agree with him in that the resolution of this signal intensity interface may signify successful therapy.
In the third letter, Dr. Holz presents a patient with multiple cerebral abscesses showing changing DW imaging characteristics after 11 days of antibiotic therapy. Similar to our findings, successful therapy was evidenced by decreasing diffusion restriction in the cavities (at least the frontal ones in his patient). He is correct in saying that the findings are also seen on the T1-weighted images, but we would venture that they are easier to appreciate on the DW image. Also, in accordance with the observations made by Dr. Duprez, we also believe that the signal intensity interface at the margins of the abscesses became less distinct with therapy.
We thank the authors of the three letters for bringing to our attention some problems in our article and for pointing out other DW imaging features of treated abscesses that we did not mention.
- Copyright © American Society of Neuroradiology