key: cord-318830-a3n4c5d1
authors: Brook, Olga R.; Piper, Kimberly G.; Mercado, Noe B.; Gebre, Makda S.; Barouch, Dan H.; Busman-Sahay, Kathleen; Starke, Carly E.; Estes, Jacob D.; Martinot, Amanda J.; Wrijil, Linda; Ducat, Sarah; Hecht, Jonathan L.
title: Feasibility and safety of ultrasound-guided minimally invasive autopsy in COVID-19 patients
date: 2020-09-17
journal: Abdom Radiol (NY)
DOI: 10.1007/s00261-020-02753-7
sha: 
doc_id: 318830
cord_uid: a3n4c5d1

OBJECTIVES: To determine the feasibility and safety of ultrasound-guided minimally invasive autopsy in COVID-19 patients. METHODS: 60 patients who expired between 04/22/2020–05/06/2020 due to COVID-19 were considered for inclusion in the study, based on availability of study staff. Minimally invasive ultrasound-guided autopsy was performed with 14G core biopsies through a 13G coaxial needle. The protocol required 20 cores of the liver, 30 of lung, 12 of spleen, 20 of heart, 20 of kidney, 4 of breast, 4 of testis, 2 of skeletal muscle, and 4 of fat with total of 112 cores per patient. Quality of the samples was evaluated by number, size, histology, immunohistochemistry, and in situ hybridization for COVID-19 and PCR-measured viral loads for SARS-CoV-2. RESULTS: Five (5/60, 8%) patients were included. All approached families gave their consent for the minimally invasive autopsy. All organs for biopsy were successfully targeted with ultrasound guidance obtaining all required samples, apart from 2 patients where renal samples were not obtained due to atrophic kidneys. The number, size, and weight of the tissue cores met expectation of the research group and tissue histology quality was excellent. Pathology findings were concordant with previously reported autopsy findings for COVID-19. Highest SARS-CoV-2 viral load was detected in the lung, liver, and spleen that had small to moderate amount, and low viral load in was detected in the heart in 2/5 (40%). No virus was detected in the kidney (0/3, 0%). CONCLUSIONS: Ultrasound-guided percutaneous post-mortem core biopsies can safely provide adequate tissue. Highest SARS-CoV-2 viral load was seen in the lung, followed by liver and spleen with small amount in the myocardium. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (10.1007/s00261-020-02753-7) contains supplementary material, which is available to authorized users.

In the setting of emerging infections, standard autopsy poses a risk to hospital staff and typically requires additional precautions [1] . Therefore, efforts to collect tissue for much needed research of COVID-19 have been significantly hampered [2] . Ultrasound-guided percutaneous core biopsies can provide tissue for research study in settings where biohazard precautions are difficult to implement [3] [4] [5] . Minimally invasive autopsy also limits body disfigurement so that patients' families may be more willing to consent to the procedure [6] .

Here we report our experience with large tissue volume ultrasound-guided minimally invasive autopsy for research purposes in patients who died with SARS-CoV-2 coronavirus infection. Samples were assessed for quality and quantity of tissue obtained by ultrasound-guided percutaneous core biopsies using tissue weight, number of cores, preservation of tissue morphology by histology, and viral load as the metrics of performance. The histologic findings are compared to previously reported full autopsy findings in patients with COVID-19 [7] [8] [9] [10] .

Patients' families were consented for limited autopsies by a pathologist during a witnessed phone call immediately after the death after referral from the intensive care unit (ICU) team. Research using autopsy tissue for this project was approved by institutional IRB. A HIPAA waiver was granted by the institutional review board for access to the patients' charts for each project using the tissue. Tissue was provided to research teams per previously IRB-approved research protocols.

Patients who expired in the intensive care units in our institution with SARS-CoV-2 infection between April 22 and May 6, 2020 were eligible to be included in the study. Autopsies were only performed during daytime weekday hours to accommodate staffing and contact with the family for consent. Five consecutive patients that expired in our institution's intensive care unit from COVID-19-related disease, during designated hours were included in this study.

Minimally invasive ultrasound-guided autopsy was performed within 3 h of death to maintain tissue viability.

Ultrasound-guided biopsies were performed by a boardcertified diagnostic and interventional radiologist with 7 years of post-fellowship experience (ORB), who reviewed available pre-mortem diagnostic imaging studies prior to the autopsy for alteration to general anatomy that would preclude percutaneous biopsy access, such as depth of the organ from the skin surface, lack of the organ, replacement of normal tissue etc. The procedure was performed in the morgue on the standard morgue table in the supine position. The morgue is negative pressure room, vented outside to the building roof. Transfer of the body was performed while in the body bag and made easier in some patients when the body had been brought to the morgue on a metal tray rather than a hospital bed. The body bag was not opened until all personnel were wearing their protective gear. The body was not removed from the bag, and towels/drapes were tucked under either side of the body to protect the inner side of the bag. Radiologist, pathologist, and research assistant participating in the autopsy wore standard protective equipment used at autopsy that included yellow gown, eye shield, gloves, and shoe covers with the addition of an N95 mask. All protective equipment was donned prior to opening of the body bag. After autopsy, the body was wiped of watersoluble ultrasound gel, the towels/drapes were removed, and the bag was zipped and wiped with disinfectant before being placed back into the morgue cold room. Protective equipment was then doffed.

Standard donning and doffing of the protective equipment was performed, followed by washing hands with regular soap.

During the procedure, entry to the autopsy room was restricted to the radiologist, pathologist, and technician processing the samples.

Biopsies were obtained with direct ultrasound guidance with iU22 system (Philips Medical, Shelton, CT) with convex 5-1 MHz transducer for liver, spleen, kidney, and heart and linear 12-5 MHz transducer for lung, breast, and testicle specimens (Fig. 1) . Fifteen preapproved research protocols requested a number of fresh frozen and formalin-fixed samples with total of 30 samples from the lung (10 fresh frozen and 20 in formalin), 20 samples from the liver (10 fresh frozen and 10 in formalin), 20 from the kidney (10 fresh frozen and 10 in formalin), 20 from the heart (10 fresh frozen and 10 in formalin), 12 from the spleen (5 fresh frozen and 7 in formalin), 4 from the subcutaneous fat (2 fresh frozen and 2 in formalin), 2 from the skeletal muscle (2 fresh frozen), 4 from the testis in males (2 fresh frozen and 2 in formalin), and 4 from breast tissue in females (2 fresh frozen and 2 in formalin) with total of 112 cores from each patient. Biopsies were obtained through 13G coaxial guide with 14 core biopsy Achieve system (Merit Medical, South Jordan, UT) with 20 mm sample length. Multiple core biopsies were acquired from each organ by tilting the coaxial needle a few degrees in different directions to avoid following the same path for all biopsies. The biopsies were not directed to target specific pathology.

In the lung, under direct ultrasound guidance and linear transducer, the needle tip of the coaxial needle was followed to the pleural surface in the right mid lung field from anterior mid-clavicular approach and then advanced additional 2 cm deep into the lung parenchyma. The needle could not be identified with ultrasound beyond the pleura (as expected due to comet tail/reverberation artifact caused by gas in the lung). In the liver, the coaxial needle was placed in the right lobe of the liver with trajectory avoiding large vessels and dilated biliary ducts, if present. In the spleen, the coaxial needle was directed away from the hilum with trajectory avoiding large vessels. In the heart, the coaxial needle was placed in the anterior left ventricle through the left anterior mid-clavicular or left parasternal chest wall. Renal cortex was targeted with patient in the supine position and coaxial needle coursing through the liver or intra-abdominal fat. In one patient, an obese patient was supported in a partial decubitus position to facilitate lateral access to the kidney. Skeletal muscle was obtained from the medial mid-thigh targeting quadriceps muscle. Subcutaneous abdominal fat was obtained in a superficial trajectory to avoid entering the muscles of the abdominal wall and peritoneum. The testis was targeted through the tunica into the testicular parenchyma in the longitudinal direction of the testes. Breast tissue was obtained adjacent to the nipple.

During the course of the procedure, tissue was processed by a senior board-certified pathologist (JLH; 20 years of postfellowship experience) using pre-labeled polypropylene cryovials (one core per vial) and containers of 10% buffered formalin (formalin-fixed cores were grouped in one container for each organ). Cryovials were immediately immersed in liquid nitrogen and then stored on dry ice until transfer to a -80 °C freezer for distribution to research groups. Formalin-fixed tissue was considered to be non-infectious after 24 h and was processed in the clinical lab under the autopsy accession number created for each case. Frozen tissue was considered infectious and only distributed at the direction of the hospital biosafety officer.

Immunohistochemistry, RNAscope in situ hybridization, Quantitative RT-PCR, and subgenomic mRNA assay were Transhepatic approach to the right kidney in patient 2. Tip of the needle is shown with white arrow. c Needle tip (arrow) in the right lobe of the liver, with the trajectory avoiding major vessels and large bile ducts. d Biopsy of the right lung parenchyma with the needle tip just beyond the pleura (arrow) in patient 1. e Biopsy of the spleen in patient 5 with arrow pointing to the tip of the needle. The trajectory of the biopsy is chosen to avoid hilar vessels. f Linear transducer was used to obtain the biopsy of the testes. Arrow pointing to the course of post-biopsy track in the testicular parenchyma performed to assess preservation of viral RNA and determine tissue viral burden (Online Appendix 1) accordingly to previously described methods [11] [12] [13] .

None of three participants in the autopsy has shown any sign of illness nor met institutional criteria required for COVID testing during the study period and 6 weeks after cessation of the study.

Five patients were included in the study, 2 males and 3 females, with age range from 58 to 91 years with variety of comorbidities ( Table 1 ). All patients underwent minimally invasive autopsy within 3 h of death with time between death and initiation of tissue collection at 120 min, 165 min, 150 min, 90 min, and 85 min. Limited hours of operation of this study resulted in approaching for consent 8% (5/60) of all patients that expired with COVID-19 in our institution in the study period. All families that were offered minimally invasive autopsy for their family member consented to and proceeded with the procedure.

Protocol-specified samples were obtained in almost all patients: lung (5/5, 100%), liver (5/5, 100%), spleen (5/5, 100%), kidney (3/5, 60%), heart (5/5, 100%), testes (2/2, 100%), breast (3/3, 100%), skeletal muscle (5/5, 100%), and fat (5/5, 100%), (Table 2 ). In total, in 3 patients in which all organs were successfully sampled 112 cores in each patient were obtained and in 2 patients in which kidney samples were not obtained 92 cores in each patient were obtained.

Weight of the samples for patient 3 was 15 mg for lung, 12 mg for liver, 20 mg for kidney, 13 mg for myocardium, and 18 mg for spleen. The specific tissue sampling was verified by gross examination of the cores for color and texture ( Fig. 2 ) with spleen having a dark red color, heart, liver, and kidney having a tan color, and lung being heterogeneous white and red/hemorrhagic and spongy when added to liquid. Tissues were further verified the next day by histology. Samples of non-targeted fat tissue obtained inadvertently during the intervention were excluded based on the following criteria: fat was yellow, did not hold a cylindrical shape, and was easily smeared on a paper towel. Accidental sampling of the fat would indicate unintentional needle displacement and require ultrasound-guided repositioning of the coaxial needle prior to acquisition of additional samples.

In two patients with chronic end stage renal disease (patients 1 and 4), the kidneys were not targeted due to severe atrophy or extensive cystic changes in the kidneys and the high likelihood of non-diagnostic yield. In one patient, sampling of the heart was limited, although representative, because of altered anatomy with tissue cores including portions of stomach and lung, likely due to anterior pleural adhesions and hiatal hernia. Due to large number of samples that needed to be obtained from each organ, the coaxial needle was slightly repositioned every few samples to prevent needle coursing through the same track and thereby maximize the yield of each core biopsy. Coaxial needle was utilized to prevent the need for repeated percutaneous access and targeting for each sample. Supine position of the patient made access to the kidney somewhat challenging due to its retroperitoneal location, especially in larger patients. In one patient, partial decubitus position improved ability to obtain renal samples.

Histopathological evaluation of the samples showed pathologies similar to those described in the published autopsy studies [7] [8] [9] [10] 14] , indicating that the random sampling was representative of COVID-19-related injury. Detailed description of patients' course and pathological findings are shown in Table 2 and Fig. 3 . Characteristic findings in the lung included three patterns of injury, which were present heterogeneously throughout each core and among different cores. The dominant pattern in the lungs was pulmonary edema and fresh hemorrhage within alveolar spaces and minimal interstitial inflammation or edema. The second pattern was prominent interstitial chronic inflammation and edema, alveoli with type-II pneumocyte hyperplasia and multinucleation, and the third pattern was diffuse alveolar damage with organizing fibrin in alveolar spaces, and extending from alveolar septa, representing hyaline membranes. No intravascular microthrombi were identified distinct from the alveolar fibrin (Fig. 3) . In several patients, the heart showed a pattern of viable myocardium with mild edema and a sparse lymphocytic infiltrate, predominantly around small vessels (Fig. 3) . The spleen consistently showed increased red pulp with patchy necrosis or larger infarction, possibly related to shock. White pulp was diminished and sinusoids were filled with large activated lymphocytes (Fig. 2) . The liver showed centrilobular necrosis (Fig. 3) , and the kidney showed non-specific changes of acute tubular injury consistent with shock (Fig. 3) . The histology of each patient was consistent with the clinical data (Online Appendix 2).

The viral load varied by organ ( Table 2 ). All 5 patients (100%) had high levels of virus in their lung samples. Lower levels of virus were present in liver 4/5 (80%), spleen 4/5 (80%), and heart 2/5 (40%). There was no virus detected in the kidney samples (0/3, 0%).

Immunohistochemistry and RNAscope in situ hybridization for viral nucleocapsid protein showed staining in 4 of 5 cases (case 3 was negative). Likely explanation for lack of nucleocapsid staining while having a viral load detected is due to long course of disease in this patient. This patient likely succumbed to complications of the COVID and not during initial infection. Staining was predominantly in the lung (Fig. 4) , where viral loads were also very high Chest X-ray in patient 1 obtained a day prior to death showed diffused patchy bilateral opacities and cardiac enlargement, which were new compared to X-ray a week prior (a). Chest X-ray performed 6 days prior to death showed persistent bibasilar airspace opacities (b). Similarly, chest X-ray obtained a day prior to death in patient 3 showed bilateral airspace opacities (c). Chest X-ray obtained a day prior to death in patient 4 showed extensive perihilar and lower lobe bilateral consolidations, progressed from prior (d). Chest X-ray obtained a day prior to death in patient 5 showed left lower lung consolidation and subcutaneous emphysema, s/p chest tube for right pneumothorax (e) ( Table 2) . One case showed focal staining in the sinusoids of the liver.

Immediate pre-mortem imaging was sparse, and included chest X-rays (Fig. 5 ) that showed bilateral predominantly basilar airspace opacities.

We have shown the safety and good diagnostic quality of core tissue sampling for research purposes in COVID-positive patients using a minimally invasive ultrasound-guided approach. We were able to obtain all required samples from the variety of organs within 3 h of death, with preservation of tissue quality for histology and ancillary research studies.

Our results regarding tissue quality are comparable to prior reports of minimally invasive autopsy of tumor tissues obtained under computerized tomography guidance by Van Der Linden et al. [15] . In that study, they found that tissue quality for RNA analysis of flash frozen core samples was comparable to open autopsy, and declined with post-mortem interval. Their best results were from tissue taken within 3 h of death. In our study, the time of acquisition ranged from 85 to 165 min but shortened with each consecutive patient as more experience was gained. Time to snap freezing and formalin fixation was shorter since tissue was processed during the case.

Despite limited and non-lesion-directed sampling, the histology was consistent with those described in COVID-19 patients [7] [8] [9] [10] 14] , including pulmonary edema and hemorrhage with diffuse alveolar damage, sparse lymphocytic infiltrate around small vessels in the myocardium, increased splenic red pulp with patchy necrosis, and centrilobular necrosis in the liver. The small number of patients, lack of comparison group, and presence of comorbidities makes it impossible to directly relate the findings to solely COVID-19 infection based on this small study.

Although not commonly used, minimally invasive autopsy has previously been shown to provide similar diagnostic results and tissue for ancillary testing to the traditional autopsy [4, 5, 16] . As shown in our study, minimally invasive approach appears to be a viable alternative in high-risk settings for patients with infectious diseases in absence of adequate protection to the morgue staff [1] . This approach has been used previously in a recent outbreak of Yellow Fever in Brazil [3] , as well as in prior studies from Mozambique with variety of infectious etiologies [16, 17] . One previous COVID-19-related autopsy study using ultrasoundguided core biopsies has been reported, but was limited to lungs, liver, and heart, and did not include details about the imaging or tissue volume obtained [9] .

SARS-CoV-2 viral loads of the freshly frozen tissue samples provided additional insight into the spread of the virus in the body. Highest concentrations of the virus were seen in the lungs, with small to moderate amount of virus in the liver, spleen, and heart. No virus was detected in the kidneys. This is surprising given the degree of renal failure in these patients. Viral infection of proximal tubules has been described [18] . It appears that obtaining viral loads is safe and feasible with minimally invasive biopsy from variety of organs.

Our study has a number of limitations. The number of patients included was small due to recruitment being constrained by staffing availability during a time when hospitals were at high census due to the pandemic. The urgency of providing research scientists with sufficient tissue for investigation in a safe manner during an ongoing pandemic was the main priority of this study. Interestingly, not a single patients' family declined the minimally invasive autopsy approach, potentially indicating a high level of acceptance as the procedure does not cause body disfigurement [6] . It is reassuring that the protocol in this study was found to be safe for personnel but this would need to be confirmed by larger studies. Other limitations include lack of assessment of the level of experience required to obtained sufficient quality tissue. Only a single experienced radiologist and pathologist participated in this study. It is possible that without this expertise the results would be less successful and additional training would be required.

Post-mortem ultrasound guidance has inherent limitations. First, residual air in the aerated lung does not allow tracking the needle beyond the pleura. Therefore, the needle was tracked to the pleura and then the trajectory was extrapolated for additional 2 cm into lung parenchyma. Kidney can be difficult to target from the anterior approach in the supine position due to its retroperitoneal location. Partial decubitus on the autopsy table facilitated access to the kidney. Large body habitus and pre-existing conditions (e.g., pneumothorax, small kidneys in ESRD) may complicate ultrasound guidance further, but this could be overcome with review of pre-mortem imaging, if it was obtained. Bowel and mesentery can be visualized by ultrasound, but these structures are thin and not amenable to a needle biopsy, and therefore, we have not attempted to sample these structures. Brain and blood vessels are not amenable for percutaneous ultrasound-guided biopsy, and thus, these structures needs to be evaluated by full traditional autopsy when it is required. Access to the heart was limited in one patient by overlying lung/pleural adhesions. Onsite visual evaluation of the gross core samples ensured that adequate tissue rather than fat was obtained. Lastly, major limitation of this study is small number of patients included. Future studies with larger number of patients and correlation with full traditional autopsy would be helpful to confirm whether minimally invasive autopsy provides a sufficiently representative sample.

In conclusion, we have shown that minimally invasive ultrasound-guided autopsy is feasible and safe to obtain large number of samples from multiple organs in COVID-19 patients for research purposes.

Autopsy in suspected COVID-19 cases

Autopsy slowdown hinders quest to determine how coronavirus kills

Ultrasound-guided minimally invasive autopsy as a tool for rapid post-mortem diagnosis in the

Sao Paulo yellow fever epidemic: Correlation with conventional autopsy

Multimodality minimally invasive autopsy-A feasible and accurate approach to post-mortem examination

Ultrasonographic autopsy (echopsy): A new autopsy technique

Willingness to Know the Cause of Death and Hypothetical Acceptability of the Minimally Invasive Autopsy in Six Diverse African and Asian Settings: A Mixed Methods Socio-Behavioural Study

Histopathology and Ultrastructural Findings of Fatal COVID-19 Infections

Pathological study of the 2019 novel coronavirus disease (COVID-19) through postmortem core biopsies

Pulmonary and Cardiac Pathology in Covid-19: The First Autopsy Series from New Orleans

SARS-CoV-2 infection protects against rechallenge in rhesus macaques

Defining HIV and SIV Reservoirs in Lymphoid Tissues

Virological assessment of hospitalized patients with COVID-2019

Pulmonary Vascular Endothelialitis, Thrombosis, and Angiogenesis in Covid-19

Post-mortem tissue biopsies obtained at minimally invasive autopsy: An RNA-quality analysis

Validity of a Minimally Invasive Autopsy for Cause of Death Determination in Adults in Mozambique: An Observational Study

Infectious cause of death determination using minimally invasive autopsies in developing countries

Acute Kidney Injury in COVID-19: Emerging Evidence of a Distinct Pathophysiology

Acknowledgements We are grateful to Dr Deb Burstein for her comments in reviewing the manuscript.

Conflict of interest All authors have no conflicts of interest to declare.