Dynamic Monitor of CT Scan Within Short Interval in Invasive Pulmonary Aspergillosis for Nonneutropenic Patients


 Background: In nonneutropenic patients with underlying respiratory diseases (URD), invasive pulmonary aspergillosis (IPA) is a life-threatening disease. Yet establishing early diagnosis in those patients remains quite a challenge. Methods :A retrospective series of nonneutropenic patients with probable or proven IPA were reviewed from January 2014 to May 2018 in Department of Respiratory Medicine of two Chinese hospitals. Refer to the relevant diagnostic criteria in the 《American Society of Infectious Diseases Guidelines for Invasive Aspergillus 2008》1. Those patients were suspected of IPA and underwent lung computed tomography (CT) scans twice within 5-21 days. The items required for IPA diagnosis were assessed by their host factors, mycological findings and CT scans according to EORTC/ MSG criteria. Results: Together with the risk factors, mycological findings and nonspecific radiological signs on first CT, ten patients were suspected of IPA. With the appearance of cavities on second CT scan in following days, all patients met the criteria of probable or possible IPA. Except one patient who refused antifungal treatment, nine patients received timely antifungal treatment and recovered well. One of the nine treated IPA cases was further confirmed by pathology, one was confirmed by biopys. Conclusions: Dynamic monitor of CT scan provided specific image evidences for IPA diagnosis. This novel finding might provide a noninvasive and efficient strategy in IPA diagnosis with URD.


Introduction
Invasive fungal infection refers to the fungus that grows and multiplies in the body's tissues, organs and blood after entering the body, causing tissue damage and triggering a series of in ammatory reactions. Environmental molds that are ubiquitous in the air are the main cause of invasive pulmonary aspergillosis (IPA). IPA usually affects immunocompromised individuals such as solid organ transplant recipients and patients with hematological malignancies including hematopoietic stem cell transplant recipients. Some research results show that IPA is associated with signi cant morbidity and carry a crude mortality rate of up to 30-40% in some risk groups 2, 3 Aside from those high-risk groups, the incidence of IPA in nonneutropenic patients with underlying respiratory diseases (URD) such as chronic obstructive pulmonary disease (COPD), asthma, lung cancer or autoimmune diseases with pulmonary involvement is increasing [4][5][6] .
Patients with COPD were reported to be most vulnerable for IPA development 5,7 .The mortality of IPA in URD patients has been found to be between 32% and 100% [7][8][9] . Patients with URD have similar symptoms, signs and radiology, which is likely to cause missed or misdiagnosed IPA in clinical diagnosis 3-5 . Due to the lack of speci city of the clinical manifestations of IPA, early diagnosis is di cult and the treatment effect is poor, so the mortality rate is extremely high. To establish diagnosis of IPA in URD patients without the classic risk factors is usually di cult, although several diagnostic criteria such as EORTC/ MSG Criteria 10 and Bulpa Criteria had been applied in daily practice 11 . IPA patients with URD usually present severe clinical conditions and poor lung function which make it di cult to obtain sterile lower respiratory tract samples by bronchoscopy. So sterile samples are rarely collected in daily practice, despite they are important for IPA diagnosis. Non-speci c symptoms and signs and insu cient accuracy of diagnostic tests delays early identi cation and timely antifungal treatment, thus leading to an increased physical and psychological burden.
Novel tests for diagnosis of IPA in patients with URD are under development. Next-generation sequencing (NGS), Aspergillus-speci c lateral-ow device tests, bioluminescence and small molecule imaging were reported to be helpful in diagnosis of IPA 8,12,13 . However, these novel tests need be veri ed in large population and the cost of the test are very high. A simple, noninvasive and effective diagnosis method are urgently needed, especially in developing countries.
Here by adopting a strategy of repeated CT scans within a short interval, we identi ed 10 cases of IPA in nonneutropenic patients. Our data showed that IPA had common imaging signs such as consolidation and tree-in-bud pattern in the early stage, then showed typical IPA signs such as cavitated nodules and halo signs in the following days. Dynamic CT review within a short interval provided more available evidence for EORTC/ MSG criteria. This strategy might be useful in diagnosis of IPA for non-hematologic immunocompromised patients.

Materials And Methods
We retrospectively assessed patients admitted in Department of Respiratory Medicine of First People Hospital of Yuhang and Second A liated Hospital of Zhejiang University School of Medicine respectively, between Jan 2014 and May 2018. The EORTC/ MSG criteria 10 was taken as IPA diagnostic criteria (Table   1). We added URD history as host factor according to the previous study 14 . Patients were classi ed into proven, probable or possible IPA based on host factors clinical data, mycological criteria, histopathological or cytopathological examination. Details as follows: a. Possible cases required host factors and clinical data but without Aspergillus isolation or serology. Written informed consent was obtained from each patient.
b. Probable cases require host factors, clinical data (meet one of the followings in CT: dense and well-circumscribed lesion with or without a halo sign, an air-crescent sign or a cavity), and microbiological factors (isolation of Aspergillus in LRT samples, or positive serum or bronchoalveolar lavage uid Galactomannan test (GM tests).
c. Proven IPA identi cation require histopathological or cytopathological examination of lung tissue showing Aspergillus hyphae from needle aspiration or biopsy specimen with evidence of associated tissue damage, or positive culture for Aspergillus from a sample obtained by sterile procedure from the lung.
This study was approved by institutional review board of both hospitals. All procedures performed in studies involving human participants were in accordance with the Helsinki Declaration. Written informed consent was obtained from the patient.

Results
Ten patients were diagnosed as uncertain IPA at rst when they showed poor response to broad antibiotic and/or system corticosteroid ( Table 2). Table 2 is the basic information of the patients. Gender: 8 males and 2 females; Reasons for admission: 1 patient with AECOPD, 2 patients with acute exacerbation of asthma, 7 patients with pneumonia; previous medical history: 1 patient with history of esophageal cancer, 1 patient with history of prostate cancer, 1 patient with herpes zoster infection (relapse 2 months before admission); comorbidities: 1 patients with coronary heart disease, 5 patients with hypertension, and diabetes mellitus (DM) 4 cases, 2 patients with COPD. There were 6 patients with smoking history. The rest had no history of malignant tumors, hematological malignancy or long-term use of immunosuppressive agents. None of the patients had rheumatoid arthritis treated with corticosteroids or immunosuppressive agents, nor had prostate cancer and esophageal cancer after chemotherapy.
The physical examination results show: All patients had cough and sputum, eight patients had wheezing symptoms, some of them had thick wet rales, 4 patients had dyspnea, and 5 patients had body temperature between 37.5 °C and 39.1 °C. All patients had poor response to broad-spectrum antibiotic and/or corticosteroids. Two of them were admitted to the ICU.
Sputum culture was ordered for once at least and six times at most before initiation of antifungal therapy. Sputum culture results show: Three patient's sputum culture revealed Aspergillus once or twice. One patient's sputum sample reported lamentous fungi in all six sputum samples. Six patients' culture did not reveal Aspergillus at all. In addition four patients had a positive galactomannan (GM) test in blood. No patient underwent bronchoscope to obtain lower respiratory tracts samples.
CT scan results show: All ten patients underwent chest CT scan twice. A radiologist was invited to review the rst CT signs in a single-blind way (Table 3). At the rst CT scan, patients showed common signs of in ammation, like scattered peribronchial consolidations, thickening of bilateral lung texture, small nodular lesions along the bronchial tree, and the 'tree-in bud' pattern. Second lung CT scans were ordered at short intervals of 5-22 days (averaging 9.7 days).
All patients had deterioration of lesions with several nodules and cavities. The changes of lung CT in 3 patients are shown in the gure 1. The intervals of CT scan were 5, 8 and 10 days in three cases respectively.
Diagnosis results: visualized by images, majority of walls of cavities were thin. One patient had pleural wedge shape, and one developed pneumothorax. With the typical CT signs of IPA such as cavitied nodular and halo signs appearing on second CT scan, eight patients met the criteria of probable IPA, and two patients met the criteria of the possible IPA (Table 4). Only one patient had large peripheral nodules which allowed a biopsy through CT guidance. The lung tissue revealed Aspergillus ( Figure 2).
Treatment process and results ( Table 4): Eight of the nine patients were treated with voriconazole for 15 days to 6 months. One patient was treated with voriconazole at rst, but had no response. So eight days later the patient was treated with posazonazole instead and had good response. Among the nine patients who received treatment, eight of them recovered well through evaluation of symptoms and CT scan signs and survived, the survival rate was 88.89%. One patient recovered after 2 weeks of voriconazole therapy, but voriconazole was discontinued because of economic cost of voriconazole. The patient died after discharge. Only one patient refused antifungal therapy and was lost during following visit.

Discussion
Data from a German study shows, during the period from 1979 to 1992, the incidence of invasive mycosis increased by about 8 times, and IPA, as the most harmful type and the most fatal type of pulmonary aspergillosis infection, has gradually been paid attention by clinical researchers 14 . Clinically, IPA is generally divided into neutropenia and non-neutropenia. This study mainly discusses the diagnostic methods of IPA patients with non-neutropenia. We found that IPA in nonneutropenic patients showed a speci c progressive deterioration in the CT scan in a short interval, which promoted early diagnosis and timely antifungal therapy. The diagnosis of IPA was validated nally by therapeutic response and/or biopsy. Therefore, our ndings provide a noninvasive, feasible and effective strategy for early diagnosis of IPA with URD. To the best of our knowledge, the current report is the rst to emphasize the diagnostic value of dynamic monitor of CT scans in IPA with URD.
Research data shows that IPA is commonly diagnosed in neutropenic patients, but also could be diagnosed in nonneutropenic patients with URD 5,15 . In our data, all patients had no neutropenia but less severe forms of immunocompromise in lungs. Most of the reasons for their admission are COPD, asthma and pneumonia, complicated with prostate cancer, hospital acquired pneumonia (HAP), and most of them have underlying diseases such as coronary heart disease, hypertension, and diabetes. Some patients have a history of malignant tumors. So we suggest to keep IPA in mind when managing the patients in Department of Respiratory Medicine. URD of those patients could further increase the complexity of Aspergillus diagnosis. First, pulmonary diseases commonly share same symptoms and signs with IPA, such as fever, dyspnea, chest tightness, wheezing and sputum production. And those non-speci c symptoms and signs could mask Aspergillus infections. Second, corticosteroids and broad-spectrum antibiotics are used commonly even overused in these population, which could increase risk for IPA 4,5 . Third, biomarkers and speci c CT signs of IPA are not sensitive in nonneutropenic patients. The speci c CT signs like cavity or air crescent are less common in nonneutropenic patients than in neutropenic patients 8 . In agreement with another report 16 , the most common CT nding was consolidation. Such CT signs are non-speci c and might correspond to a wide range of morbidities such as bacterial pulmonary infection, cardiac failure, aspiration pneumonia and so on. At last patients' poor clinical conditions like weakness, dyspnea, hypoxic respiratory failure and cardiac failure made invasive procedures such as lung tissue biopsy and bronchoscopy risky. As in our report, only one patient received biopsy. Yet tissue biopsy and lower respiratory tract samples for culture or GM tests in BALF are very speci c for IPA 10,17 .
GM is a universal polysaccharide component in the cell wall of Aspergillus, which is a polyantigen. GM appears in circulation about 1 week earlier than clinical symptoms and imaging abnormalities. Continuous monitoring of patients' serum GM levels is helpful for early diagnosis of IPA and timely medication. And the detection of the GM antigen in BALF and serum serves as a reliable assay for the diagnosis of IPA 16 . Positive GM test has been taken as an important criterion for the diagnosis of IPA both by the EORTC/MSG and Bulpa criteria. In our report, only four out of nine patients reported positive GM test. So our results showed that GM assay have relatively low sensitivity in nonneutropenic patients, as reported previously 18,19 . Meanwhile, there were other factors affecting result of GM. One of the GM positive patients had been administered piperacillin-tazobactam prior to the test, which was reported to be one of the reasons for false positives in the serum-GM assay 20,21 .Some studies have reported that the BALF-GM assay is more sensitive than the serum-GM assay and fungal cultures 18,20,22 . This is a shortcoming that BAL procedure was not conducted through bronchoscope as common in our study. A number of reasons prevented doctors to successfully obtain BALF. First, the bad general condition of patients, extreme discomfort and side effects of bronchoscope reduced patients' compliance. Second proper standardization techniques of BAL procedure are still lacking. There were variations in the BALF volumes and GM cut-off values reported in different studies. At last the yield of BALF-GM is associated with the lavage site. Therefore, how to accurately locate the lesion is critical yet very di cult.
Previously studies reported that some special signs in CT are highly suggestive of IPA, like cavity, vessel occlusion signs 23 , patchiness 24 , airway-invasive features in nonneutropenic cases 25 . But several papers have reported that CT signs in nonneutropenic IPA is nonspeci c. So the imaging ndings of nonneutropenic IPA need further study. Secondly, given IPA was an infectious disease, it might evolve over several phases which might begin with colonization, progress to infection and, nally lead to manifestation of disease symptoms in patients 28 . This phase evolving was reported in a female case of invasive tracheobronchial Aspergillosis, in which the CT scan was order on day 1, day 4, day 7, day 21, day 63 and day 139. It was found that invasive tracheobronchial Aspergillosis could progress to IPA with extended parenchymal lesions within a short period 29 . In summary, the CT signs in nonneutropenic IPA might change over time, and speci c signs could appear in one certain time point. As showed in our report, IPA underwent a progress beginning with nonspeci c CT signs, then developing to cavities within a short period of about 9 days, which was reported as appearing 2 weeks in neutropenic IPA 30 . Until now, this is the rst report about the progress deterioration of CT scans in nonneutropenic IPA, the exact dynamic changes of CT scans in nonneutropenic IPA is far from clear, so speci c study designed to observe CT signs at different stages of IPA is warranted.
Right now there are several guidelines of diagnosis and treatment for IPA released by several committees, namely EORTC/ MSG criteria 10 , Bulpa criteria 31 and intensive care unit (ICU) criteria 31 [32].The scope for each guideline are different. EORTC/ MSG criteria is limited for cancer patients but are also widely used in other patients. The Bulpa criteria is proposed to diagnose IPA speci cally in COPD patients. The ICU criteria are proposed to diagnose IPA in the ICU setting.
Items required for proven IPA are the same in the three sets of criteria, yet the items required for probable IPA are different. Here we used EORTC/ MSG criteria to diagnose IPA. When patients had a history of severe COPD, Bulpa criteria were also used. As we found, EORTC/ MSG criteria has strict requirements regarding the typical CT ndings. So according to EORTC/ MSG criteria, probable/putative IPA should meet one of three CT signs in clinical data as follows, a). Dense, well-circumscribed lesion(s) with or without a halo sign. b) An air-crescent sign. c) A cavity. Yet those typical CT signs for IPA (e.g. halo or aircrescent sign) are particularly rare in early stages in nonneutropenic patient. As showed in our study, the rst CT scan only had some nonspeci c CT signs as reported before 3,9,32 , which were not helpful for early diagnosis and timely treatment.
Meanwhile, we found there was no requirement for dynamic changes of clinical exacerbation, neither the CT scan nor mycological ndings in the EORTC/ MSG criteria. We speculated that it was because of EORTC/ MSG criteria mainly serving for cancer or hematopoietic malignancies, which might deteriorate in hours and days. Yet in IPA in nonneutropenic patients with local airway impaired immunity, the clinical process is not usually so urgent.
Nousheen and colleagues reported the average length of hospital stay were 10.61±9.08 days 33 , and ours were 45.3 days. We found there was a very signi cant CT sign deterioration among those patients after average intervals of 9 days, at least 5 days. Our results suggested that EORTC/ MSG criteria were not sensitive enough for nonneutropenic IPA without reexamination of CT scans. Thus, the procedure of applying dynamic monitor of clinical or dynamic CT scans is a way to optimize the EORTC/ MSG criteria.

Conclusions
IPA in nonneutropenic patients with URD has become a challenge in clinical practice. By dynamically monitoring disease progression via CT, it might improve the accuracy of diagnosis, especially in seriously-ill patients who could not stand bronchoscopy and lacking positive mycological ndings. We suggest the interval of CT scans could be around a week, or at least 5 days in emergency situation based on our data. Our novel nding might provide a valuable noninvasive and e cacious strategy in nonneutropenic IPA. *, URD was added as host factors in our study according to previous report  Figure 1 Dynamic monitoring of CT scans at short intervals of 5, 8 and 10 days in three cases respectively. The rst CT scan of patients showed Figure 2