Response: Magnetic resonance lymphangiography: How to prove it?

Thank you for providing us the opportunity to respond to the comments made regarding our manuscript that presents our approach to MR lymphangiography (MRL). First, it is a bit perplexing that the author of the rebuttal believed our manuscript was based only on four consecutive patients. The manuscript title and objectives present this as a Technical Review and not a research study, and the manuscript figures show MRL images from more than four different patients. As described on pages 6 and 7 of our manuscript, the occurrence of venous uptake of intracutaneously administered gadolinium-based MR contrast is well known, and the presence of contrast within superficial veins can confound the interpretation of an MR lymphangiogram. One of the purposes of our manuscript was to explain the imaging features and techniques that we have found helpful in differentiating superficial veins from enhancing lymphatic channels. This includes differences in morphology, time course of enhancement, and the inclusion of a MR venogram as part of the MRL exam. Our surgeons routinely use ICG lymphography (ICGL) in the operating room before a lymphaticovenular bypass procedure, and have more than 5 years of experience with this lymphatic imaging technique. We are fortunate to have a close working relationship with our colleagues in plastic surgery, which has been extremely important in the development of our MRL exam protocol and our understanding of the imaging findings. Thus, we have no criticisms of ICGL. In the manuscript, we compare the advantages and disadvantages of the different imaging modalities that are clinically available to us for preoperative planning to explain why we selected MRL for this purpose. While the references in the letter appear to be mis-cited, the letter describes two “recent” articles that report “even more” advantages of ICG lymphography. The first article selected by the author presents the results of a small research study on 18 patients that correlate the findings of ICGL with ultrasound elastography. In this manuscript, Hayashi et al included three sentences in the discussion that describe advantages of ICGL that we agree with. They explain that the technique is “safe,” “minimally invasive,” “simple,” “useful to find lymphatic vessels during LVA,” and can show “lymph pump function and circulation.” The second referenced article is an invited review that describes that the ICGL findings in lymphedematous limbs can be classified into two main patterns: linear and dermal backflow. An additional advantage described in this second article is that ICG patterns can be used to stage lymphatic drainage and guide the timing of treatment. These two articles also present several disadvantages of ICGL that we had not included in our manuscript, and include the following. ICGL is not available at all institutions and is contraindicated in patients with an iodine allergy. With the ICGL technique presented in the two references the fluorescent images are obtained at two time points: immediately after ICG administration and 12– 24 h later to allow dissemination of the ICG. The inclusion of a second delayed scan creates a challenging and inconvenient workflow. In contrast, with current MRL techniques, the exam is only performed once. In addition, with fluorescent imaging, lymphatic channels can be obscured by overlying fascia, muscle, or bone. Because MRL is a three-dimensional imaging technique, overlying tissue does not hinder the visualization of lymphatic channels. Furthermore, intravascular contamination can occur with ICG administration and, thus, is not unique to MRL. Contrary to the letter author’s report that ICGL is painless, our surgeons indicate that patients do feel pain with the ICG injections, which has also been described by Narushima et al. For our MRL exams, we mix a local anesthetic with the MR contrast before the intracutaneous administration to reduce the pain felt at the injection site. We began clinical MR lymphangiography in 2011, and since then, the service has grown each year to where we currently perform five to eight exams per month. Our surgeons have found that the MRL results correlate very well with the findings on intraoperative ICG lymphography. Thus, they have come to rely on MRL to preoperatively select patients with abnormal lymphatic drainage and identifiable lymphatic channels suitable to perform the lymphaticovenular anastomoses. This would not be the case if we were not correctly identifying lymphatic channels with MRL. Thus, our clinical experience combined with the current body of literature supports that MRL does visualize abnormal peripheral lymphatic channels. The letter author alludes to criticisms raised against MRL regarding the differentiation of lymphatic channels from veins at the Lymphoedema Congresses held in 2013 and 2014. While we do not have details of the imaging techniques used by these sites, it may be possible that differences in MR hardware or imaging techniques could explain the difficulty that some sites may have in differentiating lymphatic channels from veins. Of interest, a recent study showed the potential of using an iron-based blood pool MR contrast agent to selectively suppress venous enhancement and create a dark-blood lymphatic only MRL image. As has been described, normal lymphatic channels in nonedematous limbs can be faint or not visualized with current MRL techniques. This is believed to be due to rapid lymph transport, smaller lymph volume, and smaller size of normal lymphatic channels when compared with the abnormal lymphatics in the lymphedematous extremity. The letter author should clarify the ICG pattern present in the lymphograms of the right and left limbs presented in the figure included with the letter. Since a linear ICGL pattern can be seen in a normal limb with normal lymphatic drainage, the expected finding in a normal limb would be the visualization of a normal ICGL pattern with the absence of enhancing lymphatic channels on MRL. We believe that the inability to depict normal lymphatic channels on MRL is actually an advantage