Discussion
Although for a long time it was thought that LUS was not feasible due to the air content, numerous studies in adults and subsequently in children, showed its efficacy for the diagnosis of pneumonia with sensitivity and specificity superior to CXR. (17,19-24) Therefore, LUS is now identified as a valid substitute for CXR in the course of respiratory tract infections in children with the aim of reducing exposure to ionizing radiation. Moreover, CXR does not allow localization of infection in cancer patients in over 44% of cases. (13) As demonstrated by Gerristen et al., the sensitivity of CT at low doses of radiation is 73% performed on the first day of febrile neutropenia versus the sensitivity of 36% of CXR. (13) Heussel et al. showed that in more than 50% of febrile neutropenic patients with normal CXR, there were signs of pulmonary inflammation on CT. (25)
Ultrasonography is the ideal tool for its speed, non-invasiveness, easy repeatability and simple interpretation of the examination after appropriate training even by non-radiology specialists. This last feature is of increasing importance in relation to the concept of personalized medicine. (5) The ”bed side” ultrasonography, performed in the emergency departments or in critically patients, allows a more detailed evaluation of the individual patient obtaining an ever more patient-based rather than disease-based care approach. (5)
If in pediatric clinical practice, LUS is becoming an increasingly useful examination; howevwer, no study about its accuracy has ever been performed on the pediatric cancer population, where factors such as type of neoplasia, chemotherapy and thoracic radiotherapy with its pulmonary toxic effects and neutropenia may reduce the specificity and sensitivity of the tests.
Our data indicated that the underlying cancer, the administered therapies, the number of neutrophils did not influence the result, and the specificity was 95%, comparable to that described in the pediatric population for the diagnosis of pneumonia.(26) The sensitivity, even if calculated in a small series, was 100%.
Our data, with the limit of a small sample, showed that in non-neutropenic febrile patients, LUS has an important role, comparable with the literature on pneumonia cases in pediatric age. (26-30) In this population, LUS showed better diagnostic accuracy than CXR, which had a specificity of 80%, and a sensitivity of 50%, considerably lower than literature data on pediatric population but similar to adult cancer population. (13) LUS showed to have early positivization, even before the onset of clinical respiratory symptoms. This test could therefore be used as a screening tool for pulmonary infections in febrile cancer patients, in order to undertake early specific and targeted treatments or to direct the diagnostic-therapeutic work-up. Moreover, a LUS follow-up allowed to modify or to stop the anti-infection treatment according to the evolution of the process.
In febrile neutropenic patients, despite the absence of respiratory symptoms, it was possible to find lung lesions on ultrasound in 50% of cases, indicating that this method can be an aid technique for the diagnosis of infections during neutropenia, while CXR, with its low sensitivity is not indicated in the suspicion of lung infection during neutropenia. Five out of 10 patients presented a positive LUS for pneumonia. These data show that in 50% of patients with febrile neutropenia there is an ongoing lung infection; the latter often resolves thanks to empirical broad-spectrum antibiotic therapy but in a smaller percentage of cases it can evolve towards acute respiratory failure or sepsis. (31) Likely the reason why LUS has greater sensitivity than CXR in this population can be explained by the results obtained by Shah VP et al. (32) This study demonstrated that while for pulmonary thickenings of size greater than 1 cm the CXR and the LUS have a similar diagnostic accuracy, in pulmonary thickenings smaller than 1 cm LUS has a clearly superior sensitivity. Therefore, it can be hypothesized that in the patient with a low number of neutrophils and consequent deficient inflammatory response, the infection predisposes to develop subcentimetric thickening, not detectable on CXR. This hypothesis is confirmed by the few data we collected: in 4 out of 5 febrile neutropenic patients (80%), pulmonary thickening on LUS were subcentimetric. This finding could open the door to a new simple diagnostic tool to be used in all neutropenic febrile cancer patients.
Most patients were compliant to LUS, demonstrating the low invasiveness of the method even in this population. The less cooperative patients were younger than 6 year or patients at the onset of cancer, often frightened by many diagnostic procedures (bone marrow aspirate, lumbar puncture, placement of central venous catheters or excisional biopsy), the hospital environment and the disease itself. Moreover, LUS was efficaciously performed, without the need of sedation, what is not always feasible for CXR or CT.
The possibility of using a tool, such as LUS, which is minimally invasive, repeatable, bedside suitable, radiation free and, at the same time, accurate, as a screening for the early detection of lung infection and monitoring its evolution, can have a great impact on the management and outcome of the pediatric cancer patient.
We hypothesize a role for LUS, as a first imaging technique, and propose a possible diagnostic work-up flow chart that includes LUS for the management of patients with signs of infection, with the aim of reducing radiation exposure and inserting a non-invasive and bed side examination (Fig. 2). Obviously, the proposed flow chart does not exempt from individual evaluation and specific clinical management.