thoracic masses treated with adiology percutaneous ... · endobronchial cryotherapy has been...

Hongwu Wang, MD, PhDPeter J. Littrup, MDYunyou Duan, MDYanqun Zhang, MDHuasong Feng, MDZhoushan Nie, MD

Published online10.1148/radiol.2351030747

Radiology 2005; 235:289–298

Abbreviations:KPS � Karnofsky Performance ScalePCT � percutaneous cryotherapy

1 From the Tumor Targeted Cryother-apy Center, PLA General Navy Hospital,Beijing, China (H.W., Y.D., Y.Z., H.F.,Z.N.); and Department of Radiology,Wayne State University School of Med-icine, Detroit, Mich (P.J.L.). ReceivedMay 12, 2003; revision requested July21; final revision received May 7, 2004;accepted June 28. Address correspon-dence to P.J.L., Department of Radiol-ogy, Harper University Hospital, 3990John R St, Detroit, MI 48201 (e-mail:[email protected]).

Authors stated no financial relation-ship to disclose.

Author contributions:Guarantors of integrity of entire study,H.W., P.J.L.; study concepts and de-sign, H.W., P.J.L., Y.D.; literature re-search, P.J.L.; clinical studies, H.W.,Y.Z., H.F., Z.N.; data acquisition,H.W., Y.Z., H.F., Z.N.; data analysis/interpretation, H.W., P.J.L.; statisticalanalysis, H.W., P.J.L.; manuscript prep-aration, definition of intellectual con-tent, editing, revision/review, and finalversion approval, P.J.L., H.W.© RSNA, 2005

Thoracic Masses Treated withPercutaneous Cryotherapy:Initial Experience with Morethan 200 Procedures1

PURPOSE: To perform and report initial experience with percutaneous cryotherapy(PCT) of the thorax.

MATERIALS AND METHODS: A human investigation committee approved thestudy protocol, and all patients gave informed consent. One hundred eighty-sevenpatients who were not surgical candidates underwent computed tomography (CT)-guided PCT for treatment of thoracic cancer masses. CT-visualized low-attenuating iceformation after PCT was compared with initial tumor size and location. At 1 week andat 1, 3, 6, and 12 months after PCT, the various findings seen on available CT scans andany complications were noted. �2 and Student t tests were used to identify significantdifferences in frequencies and mean values of imaging observations, respectively.

RESULTS: Ice formation was identified at CT as reduced attenuation values (inHounsfield units) within soft-tissue masses, the mean sizes of which were 4.3 cm � 0.2(standard deviation) in peripheral locations and 6.4 cm � 0.3 in central locations.Tumor size and location were independent predictors of tumor coverage by low-attenuating ice: Mean coverage was 99% for peripheral masses 4 cm or smaller (n �101) and 80% for central masses larger than 4 cm (n � 58) (P � .001). An area ofnecrotic cavitation larger than the original mass developed in 80% (77 of 96) of masseswithin 1 week and was nearly resolved by 3 months in 7% (five of 76) of masses. By 6months, minimal pulmonary scarring was noted in 56 patients and 86% of massesshowed reduced or stable size. The overall rate of pneumothorax was only 12% (22 of187 patients), and other side effects appeared to be self limited. No major bleeding orbronchial damage was noted. Two deaths in debilitated patients were temporallyrelated, and two complications involved brachial and recurrent laryngeal nerve damage.The patient with laryngeal nerve damage regained speech within 2 months.

CONCLUSION: CT-guided PCT yielded low procedural morbidity given the extentof freezing, even near mediastinal structures. Ongoing advances in cryotechnology,imaging guidance, and treatment planning may help to avoid the degree ofundertreatment of larger central masses observed in this study.© RSNA, 2005

Many radiologists regard cryotherapy as an intraoperative modality that is used mainly byurologists and general surgeons for prostate gland and liver tumor ablations, respectively (1,2).For a long time, however, cryotherapy performed by using rigid bronchoscopes has been usedextensively to treat endobronchial neoplasms and has yielded minimal morbidity (3–5).Outpatient acceptance of cryotherapy performed with flexible cryoprobes, lower cryotherapyequipment costs, and no bronchial perforation, as compared with the bronchial damageassociated with the use of lasers (ie, for Nd:YAG or photodynamic therapy), have been notedby European investigators (6). Endobronchial cryotherapy has been considered the treatmentof choice for early superficial bronchial carcinoma (7). The unique safety feature of cryother-apy—that of enabling preservation of the collagenous architecture (8)—and the resultant

Vascular and Interventional Radiology

289

Ra

dio

logy

integrity of the tracheobronchial tree (3–7)have appeared to be all but forgotten sinceheat-based therapies became more popularfor percutaneous procedures.

The use of percutaneous tumor ablationhas rapidly advanced, and radiofrequencyis commonly used at many organ sites forapplications that include treatment of pri-mary lung cancer (9) and metastases origi-nating from renal cell carcinoma (10). Ex-perience with radiofrequency treatmentsremains in the early stages, but suggesteddrawbacks include heightened pain nearthe pleura and the chest wall; necrotic fis-tula formation—with potentially pro-tracted pneumothoraces—caused by abla-tion near bronchi, and brisk scarring(cicatrization) seen at longer-term fol-low-up (11). Thus, for central tumors, anyheat-based therapy may have limited effec-tiveness owing to bronchial disruption orperforation from denatured protein. Imag-ing performed in a large population treatedwith percutaneous cryotherapy (PCT) isneeded to define the general appearance ofthe treated area and to determine the fea-sibility of the procedure, particularly forthe treatment of central neoplasms, in-cluding those with associated superiorvena cava syndrome.

There has been a resurgence of the use ofPCT for the treatment of prostate cancer inthe urology community, primarily becauseof the improved ease of use of the proce-dure and reported long-term tumor con-trol rates that are comparable to thoseachieved with surgery and radiation ther-apy (1). After the feasibility and excellentcomputed tomographic (CT) visualizationof low-attenuating ice coverage of massesin pig livers (12) were documented, a groupof interventionists (including P.J.L.) in col-laboration with Chinese investigators per-formed percutaneous hepatic cryotherapyin humans in October 1999. China alsohas a burgeoning population of individu-als with lung cancer, which leads to ap-proximately 300 000 deaths per year inthat country (13). Interventionists inChina rapidly began performing PCT ofthe thorax after the U.S. experience withCT-guided PCT in September 2000 (14).The purpose of our study was to report ourinitial experience with PCT of the thorax.

MATERIALS AND METHODS

Patients

A human use protocol was approved bythe senior medical oversight board (ie, hu-man investigation committee) of PLA Gen-eral Navy Hospital. An informed consent

form was signed by all patients. The in-cluded patients primarily were consideredto be ineligible for surgery, or they hadundergone other lung cancer therapiesthat were unsuccessful. These patients ful-filled at least one of the following inclusioncriteria: They had (a) single or multiple pe-ripheral lung masses larger than 1.0 cm indiameter, with previous therapies (ie, radi-ation therapy, chemotherapy, and/or sur-gery) having failed; (b) a nonresectable (asdetermined, by means of surgical consulta-tion, according to either tumor size andstage or health status) central lung cancer;(c) fewer than four metastases to the lungwith controlled primary cancer; (d) a massor adenopathy involving the mediastinumand/or the pericardium without distantmetastases; and/or (e) malignant pleural ef-fusion, which was considered as inclusioncriteria only when it was associated with adistinct measurable primary lung tumor.

Exclusion criteria were as follows: five ormore diffuse or bilateral masses; diffusepleural metastases with a large effusionand no measurable primary mass; centralmasses to which percutaneous access wasdifficult (eg, associated with potentialpuncture of the major vasculature en routeto a mass); severe pulmonary dysfunctionassociated with a maximum voluntaryventilation capacity lower than 39% (ie, arelative estimate of ventilatory capacitywithin 1 minute compared with the timeframe for the predicted capacity; this pa-rameter was used in lieu of the commonmeasurement of a forced expiratory vol-ume in 1 second of less than 1 L because italso yielded some insight with regard to

the musculoskeletal respiratory capacityand the degree of patient cooperation), theinability to lie flat, or respiratory distress atrest; severe cough, current respiratory dis-tress, and/or difficulty to cooperate; and/orany bleeding dyscrasias that were not eas-ily corrected.

One hundred eighty-seven patientswith 234 masses underwent 217 PCT ses-sions as inpatients between August 2001and September 2002 (Table 1). Male pa-tients (n � 137) were predominant in thisseries. The mean age of all patients was61 years (range, 41–83 years). Primarylung cancer accounted for 84% (196 of234) of the masses and 88% (165 of 187)of the patients. Tumor stage was based onCT appearance alone, as determined bythe primary treating radiologists (Y.Z,H.F., Z.N.) in conjunction with the staffoncologic pulmonologists (H.W., Y.D.).Of the 165 patients with primary lungcancer, five had stage I; 17, stage II; 20,stage IIIA; 60, stage IIIB; and 63, stage IVcancer. Of the patients with advancedstage disease (ie, stage III–IV), 89% (127of 143) had already undergone surgery,chemotherapy, or radiation therapy. Onehundred seventy-eight (76%) masseswere treated during single cryotherapysessions. Patients underwent repeated ab-lations for treatment of larger centralmasses or multiple pulmonary masses.For bilateral masses, repeated ablationswere performed during separate sessions,whereas for multiple masses in the samelung, repeated ablations were performedsimultaneously, when possible.

TABLE 1Histologic Distribution of Primary Tumors and Metastases

Cancer Type No. of Patients No. of Lesions

Primary tumorSquamous 74 (40) 77 (33)Adenocarcinoma 72 (38) 96 (41)Small cell 13 (7) 17 (7)Unknown 6 (3) 6 (3)

Total 165 (88) 196 (84)Metastasis

Lung* 10 (5) 14 (6)Colorectal 3 (2) 4 (2)Kidney 1 (.5) 2 (1)Hepatoma 1 (.5) 2 (1)Osteosarcoma 2 (1) 6 (3)Breast 1 (.5) 4 (2)Nasopharyngeal 1 (.5) 3 (1)Esophageal 1 (.5) 1 (.4)Thyroid 1 (.5) 1 (.4)Brain 1 (.5) 1 (.4)

Total 22 (12) 38 (16)

Note.—Numbers in parentheses are percentages based on a total of 187 patients or 234 lesions.* Metastasis to lung identified postoperatively.

290 � Radiology � April 2005 Wang et al

Ra

dio

logy

PCT with CT Guidance andCalculated Cryoprobe Placement

All PCT procedures were performed byusing an Elscint CT Twin II unit (Elscint,Haifa, Israel). Tumor location and sizewere noted at preoperative evaluation. Atumor was considered to be a centralmass if it abutted any of the mediastinalor hilar structures. Tumor measurementswere obtained in two dimensions on CTimages, at the level of the tumor’s mostprominent appearance in the transverseplane. Patients were placed feet first intothe CT gantry to allow easier access forsubsequent needle and cryoprobe place-ments. CT fluoroscopy was not per-formed. Therefore, the procedures in-volved direct axial placement for optimalneedle course visualization. Most pa-tients received only local anestheticagents and no sedatives. Those patientswho were sedated received only a mildsedative (morphine 2–4 mg with or with-out midazolam hydrochloride 1–2 mg).Two interventional radiologists (P.J.L.,Y.Z.) with many years of experience per-forming percutaneous biopsy and morethan 10 years of experience performingclinical cryotherapy in other organ sys-tems (P.J.L.) worked in conjunction withoncologic pulmonologists (H.W., Y.D.).

At the time of this study, only straight-shafted 3-mm-diameter cryoprobes were

available, and they were placed by usingthe technique shown in Figure 1. In ad-dition, direct axial cryoprobe placementsprecluded the CT gantry clearance of thestraight cryoprobes; thus, no CT imagingwas performed during freezing. After thetumor was localized, the treatment areawas prepared and draped in a sterile man-ner. Two percent lidocaine was injectedinto the pleural surface. To minimize thenumber of cryoprobe punctures, single-needle placement through the tumorcenter was chosen for smaller masses (ie,those 4 cm or smaller). A 19-gauge needlespecially adapted by the lead author(H.W.) to be 20 cm in length and to havea removable hub but no inner stylet wasused. The needle tip was placed throughthe mass to the far margin; then the hubwas removed. We carefully noted the pre-cise measurement of the final needledepth to determine the depth needed forsubsequent sheath and cryoprobe place-ments.

The dilator and 11-F sheath were thenslid over the needle to the depth of theneedle insertion. The needle and dilatorwere removed, leaving the sheath inplace to receive the 3-mm cryoprobe (Fig1). The cryoprobe was blindly placed tothe same predetermined depth at the fartumor margin since CT scanning couldnot be performed during freezing because

of the straight cryoprobe handles. Thesheath was then retracted approximately4 cm to expose the length at the tip of thecryoprobe that generates ice. If a substan-tial pneumothorax—that is, one greaterthan 2 cm in radial depth—was encoun-tered during initial needle placement, asmall (8-F) catheter was used for evacua-tion during PCT. If no persistent pneu-mothorax was seen, the catheter was re-moved at the end of PCT.

The cryotherapy equipment that weused consisted of an argon:helium gas–based system (CryoCare; Endocare, Irvine,Calif) and 2- and 3-mm-diameter cryo-probes (Endocare). With use of these cryo-probes, the diameters of the resulting iceballs may vary according to tissue typeand blood supply, but the freeze lengthsare consistent for both probes: approxi-mately 4.5 cm. Ice ball diameter refers tothe outer 0°C margin of the visible low-attenuating ice, but cytotoxic ice temper-atures (�20°C) are known to occur 3–5mm inside of visualized ice margins (1,2,8).For soft tissues (eg, liver and kidney), theestimated cytotoxic ice diameters that canbe achieved by using 2- and 3-mm cryo-probes are 2.0 and 2.5 cm, respectively.Two-millimeter-diameter cryoprobes havea sharp tip and were used only when adirect-stick approach was chosen insteadof the Seldinger approach described ear-

Figure 1. Three major steps to performing thoracic PCT in the current study. Left: First, needles (arrows) are placed in thetumor with CT guidance, and the hub is removed. Middle: The dilator-sheath combination (arrows) is placed, and its positionis checked on CT images. Right: The cryoprobes (arrows) are then blindly placed to predetermined tumor depths, and thefreeze cycle is begun: 20 minutes of freezing, 10 minutes of thawing, and then 20 minutes of refreezing.

Volume 235 � Number 1 Percutaneous Cryotherapy for Thoracic Masses � 291

Ra

dio

logy

lier. Aerated lung parenchyma may serveas an insulator and thus lead to larger iceballs compared with those in soft tissues.

Since no animal model with comparablylarge lung tumors exists, the following as-sumptions were made after observing thelow-attenuating ice that formed after PCTin the first several patients: For a masswithin lung tissue, a 2-mm cryoprobe mayproduce an ice ball up to 3 cm in diameter,whereas a 3-mm cryoprobe may producean ice ball up to 4 cm in diameter. Tumors4 cm or smaller were treated by using one3-mm cryoprobe. To avoid perfusion-me-diated heating, we made no attempt toplace additional cryoprobes for treatmentof central masses.

For this study, maximal freeze and thawtimes were chosen to produce the largestfeasible ice volume and ablation zone withthe available cryoprobes, especially sinceimaging could not be performed duringthe procedure. To minimize pulmonarypunctures, no thermocouples were used tomonitor the temperature near tumor mar-gins (similar to protocol used to treat pros-tate masses [1]). A treatment cycle con-sisted of a 20-minute freezing followed by a10-minute thawing and then a 20-minuterefreezing. The thaw cycle was defined asthe cycle during which the temperaturewithin the cryoprobe (ie, obtained from athermocouple within the cryoprobe andapproximating central ice ball tempera-tures) was allowed to increase either to 0°–5°C or until 10 minutes passed. With thisprotocol, we attempted to ensure maximalthawing of the ice ball and cytotoxicity ofthe ablation zone due to osmotic shock atthe treatment margins. This thoroughthawing phase is different from commonlyused thawing phases (1,2) in which cryo-probes are not allowed to become mobilebecause their temperatures are kept lowerthan �4°C. With use of the “stick” modeon the cryoprobe system, one sets the ther-mostat to briefly activate the argon gasflow and thereby prevents cryoprobemovement. With our protocol, at the endof the second 20-minute freezing, the cryo-probes are allowed to undergo active thaw-ing whereby helium is circulated throughthe cryoprobes. Owing to the isenthalpicproperties of helium (ie, low molecularweight and low latent heat from vaporiza-tion), the cryoprobes are gently heated sothat they can be removed within 2 min-utes.

Postprocedural Evaluation and CT

Immediately after cryoprobe removal,the patient could fit back into the CTgantry for scanning and measurement of

low-attenuating ice formation. All CTscans obtained after PCT were acquiredwithin the 2–3 minutes following the ter-mination of freezing; during this time,the visible ice size remains relatively un-changed (1,2,8). It takes approximately 2hours of CT time to perform most PCT

procedures, including cryoprobe place-ment and removal. CT attenuation val-ues ranging from 0 to �10 HU were usedto confirm solid ice formation. To esti-mate the percentage of any area of unfro-zen tumor rim, bidimensional ice marginmeasurements were multiplied to pro-

Figure 2. CT image findings of immediate and short-term cryotherapy-induced cavitation.(a) Left: Slightly oblique transverse image obtained during PCT shows needle placement (arrow)in a 3-cm lobulated primary lung cancer. Right: Corresponding image obtained immediately afterfreezing shows low-attenuating ice throughout the mass. The boxes posterior to the tumorcontain attenuation measurements (in Hounsfield units) that confirm the subzero temperature(ie, tissue ice formation) within the small circular regions of interest in the tumor. (b) Transversesoft-tissue (left) and lung window (right) CT images obtained 4 days after PCT show a well-circumscribed cavitary region covering the area of frozen tumor and the adjacent rim of pulmo-nary parenchyma seen in a. The treatment area appears to be more posterior now because thepatient is no longer oblique but rather is lying supine.


Ra

dio

logy

duce an area value that could be com-pared with a similar area measurement ofthe unfrozen tumor at the same ana-tomic level (eg, 2.7 � 3.0-cm ice in a3.0 � 3.0-cm tumor � 8.1 cm2/9.0 cm2 �90% ice coverage). In addition, CT scanswere obtained in patients who were avail-able at 1 week and at 1, 3, 6, and 12months after PCT. These scans werejointly evaluated by cryotherapy teammembers (H.W., Y.D.) and the primaryradiologists (P.J.L., Y.Z., H.F., Z.N.), who

had more than 10 years experience withchest CT.

The Karnofsky Performance Scale (KPS)(5) was used to rate the palliative effectsof PCT on symptoms at 1 week followingthe procedure. Explanations of the KPSnumeric scores are as follows: A score of100 indicated that the patient had nocomplaints after PCT; 90, that the patienthad minor signs of disease but could per-form normal activities; 80, that the pa-tient had some signs or symptoms of dis-

ease and could perform normal activitieswith effort; 70, that the patient couldcare for himself or herself but was unableto perform normal activities or engage inactive work; 60, that the patient requiredoccasional assistance but could attend tomost of his or her personal needs; 50,that the patient required considerable as-sistance and frequent medical care; 40,that the patient was disabled and re-quired special care and assistance; 30,that the patient was severely disabledand hospitalization was needed, butdeath was not imminent; 20, that thepatient was very ill and hospitalizationand active support treatment were neces-sary; 10, that the patient was moribund,with fatal processes progressing rapidly;and 0, that the patient died.

All feasible efforts to obtain repeat CTscans were taken. The available results ofCT examinations of peripheral and cen-tral locations, including tumor coverageby ice according to tumor size and loca-tion, related complications, and short-term clinical follow-up findings, wereevaluated over time.

Statistical Analyses

Assessments were limited to those ofobservational differences and were notintended to power the sample size of thestudy. In addition, sample size was notdetermined to address therapeutic out-comes because the follow-up period wasso short and follow-up examinationswere limited to initial radiologic imageassessments. Patients with multiple tu-mors were limited to the metastaticgroup and were not considered in sepa-rate outcome assessments. All meanvalue comparisons were performed by us-ing the two-tailed Student t test. Fre-quency comparisons (ie, percentage tu-mor coverage by ice) were performed byusing the �2 test. P � .05 indicated statis-tical significance. The lead authors(H.W., P.J.L.) performed the statisticalanalyses by using calculated fields andincorporating appropriate formulas on astandard spreadsheet (Microsoft Excel,Redmond, Wash).

RESULTS

CT Performed Immediately afterCryotherapy

Ice was well seen on the CT imagesobtained immediately after cryoprobe re-moval (Figs 2–5) as distinct lower atten-uating (40–60-HU decrease from pre-freeze attenuation value) regions withinthe soft-tissue masses (prefreeze attenua-

Figure 3. Transverse CT images obtained during cryotherapy for treatment of a small pulmo-nary mass and during follow-up. (a) Preprocedural (left) and procedural (right) images of a 2-cmprimary lung cancer in a patient who was not eligible for surgery show the initial needleplacement (arrow). The tip was subsequently advanced to the far tumor margin. (b) Left: Imageobtained 10 weeks after PCT shows resolution of the cavitary effect that developed (not shown),with a residual parenchymal reaction but a minimal underlying soft-tissue component. Right:Image obtained 6 months after PCT shows nearly complete resolution of the parenchymalreaction and minimal residual scarring.


Ra

dio

logy

tion, 30–50 HU). However, the markedlow attenuation (approximately �800HU) of the adjacent aerated lung paren-chyma made it impossible to visualizefrozen tissue in the adjacent lung imme-diately after the freezing despite multiplewindow and level adjustments. There-fore, we could not document the desiredvisualization of extended areas of freez-ing (ie, 5–10 mm beyond the tumormass) as can be done for other organs;rather, we documented only estimates ofice coverage of the soft-tissue compo-nents of the thoracic masses.

In an attempt to quantify initial treat-ment success, the estimated coverage oftumor areas by lower attenuating ice, ascalculated by using bidimensional mea-surements, is summarized in Table 2.Tumor size and location were highlysignificant independent variables fordetermining the likelihood of tumorcoverage by ice (P � .001). Tumor stageand type did not correlate with ice cov-erage once size and location were con-sidered. The percentages of masses forwhich follow-up CT data collected overtime were available are listed accordingto degrees of cavitation and changes intreated tumor size in Table 3. An area ofcavitation larger than the original tu-mor was seen in 80% of treated areas 1week after PCT (Fig 3, Table 3); a similardegree of necrotic cavitation has been

previously noted in patients with livermasses (2). The cavitation rate had de-creased to only 7% (five of 76 treatedareas) by 3 months, suggesting resorp-tion of nearly all necrotic debris. By 6months, 86% (48 of 56) of the treatedareas were stable or smaller than theoriginal tumor.

Complications and Early Follow-upData

The complication rates listed in Table 4emphasize a low pneumothorax rate anda predominance of self-limited (ie, last-ing fewer than 5 days) side effects, de-spite major central freezing. No intrapro-cedural deaths were noted, but twodeaths were temporally related: One pa-tient died of a pulmonary embolus 1 dayafter cryotherapy, and the other died ofacute respiratory distress syndrome aweek later. However, no technical diffi-culties or other complications occurredduring the procedure for either of thesepatients. Although follow-up data weretoo limited for the calculation of accuratesurvival estimates, the palliative benefitsof PCT in terms of KPS performance sta-tus and general health status (eg, in-creased dietary intake and weight gain)were noted. In the group of 143 patientswith advanced-stage cancer, general healthand KPS statuses improved significantly 1

week after cryotherapy: The mean KPSscore increased from 75.2 � 1.3 before to82.6 � 1.4 1 week after PCT (Student t �3.87, P � .01).

DISCUSSION

The short follow-up period for this largeseries of patients who underwent tho-racic PCT precluded the calculation ofreliable estimates of the effectiveness ofthis therapy for treatment of primary andmetastatic lung tumors. That said, thelong history of effective cryotherapy per-formed at other organ sites (1,2) makesthe results obtained in this series encour-aging in terms of at least patient safetyand the feasibility of PCT for many tho-racic locations. The observation that theminimal sedation used during PCT waswell accepted by conscious patients sug-gests that PCT may become a treatmentoption for patients for whom general an-esthesia poses risks, as well as those whoare not surgical candidates. Several clin-ical assumptions and technical short-comings need to be clarified since bothcryotherapy equipment and imagingtechniques continue to advance.

The diversity and large number of pa-tients who underwent PCT in this studygenerate confidence in the technical as-pects of thoracic tissue responses to cryo-

Figure 4. Transverse CT images obtained during cryotherapy for treatment of superior vena cava syndrome with use of multiple cryoprobes andduring short-term clinical follow-up in patient who presented with facial edema and prominent jugular venous distention. Because of the lack ofreadily available emergent radiation therapy in China, this patient was treated for symptomatic relief via the immediate (2–3-day) cryotherapeutic“softening” that a tumor undergoes after ablation. Left: Contrast-enhanced image shows the superior vena cava (arrow) being displaced andcompressed by bulky mediastinal and right hilar adenopathies. Middle: Nonenhanced image obtained immediately after cryoprobe removal showsthe superior vena cava (large arrow) bracketed by low-attenuating ice, with air in the central cryoprobe tracks (small arrows). The low-attenuatingice extended into the right main bronchus (arrowheads) without complications. Right: Photograph of patient 3 days after PCT shows only mildlydilated neck veins (arrows) and resolved facial edema.


Ra

dio

logy

therapy. We acknowledge the potentialbias in the selection of our patient groupsand that cryotechnology was rapidly im-proving during the course of this study.The individuals treated with PCT were

mainly patients who were ineligible forsurgery, and 89% (127 of 143) of themhad advanced-stage cancer for whichprior treatment had failed. Thus, the out-comes of these patients would be difficult

to compare because the PCT procedureswere only palliative and were used fortreatment of several tumor types. Thelarge number of procedures, however,still enabled an accurate assessment oftechnique, and the results established therelative safety and feasibility of PCT fornearly all tumor locations.

The complementary effects of tumorsize and tumor location on immediatepostfreezing outcomes helped us redefinetreatment goals and better generally plancryotherapy procedures. Newly availableangled cryoprobes allow much greaterflexibility so that the degree of tumorundertreatment observed in this study,which occurred primarily because directCT imaging could not be performed dur-ing cryoprobe placement and freezing,can be avoided.

Despite the limitations, thorough (99%)ice coverage was achieved by using a single3-mm cryoprobe for 101 peripheral masses4 cm or smaller. We now advocate brack-eting small masses with single-puncture,2.4-mm angled cryoprobes. By first placinga 20-gauge needle (eg, Westcott needle)near the center of the mass, one can di-rectly place the larger cryoprobes with atrajectory parallel to the 20-gauge needle,allowing minimal repositioning and facili-tating a lower risk of pneumothorax. How-ever, central masses larger than 4 cm mayrequire the most conservative cryoprobedistribution (eg, one cryoprobe for each1.5–2.0-cm-diameter tumor) since in ourstudy, only 80% coverage of these tumorswas achieved. The intermediate categoriesof peripheral masses larger than 4 cm andcentral masses 4 cm or smaller also suggestthe need for closer cryoprobe arrange-ments to maximize the freezing capacityper centimeter of tumor area (eg, one cryo-probe for each 2.5-cm diameter, similar tothe placement protocol for other organs).

The ability to visualize lower attenuat-ing ice as it thoroughly covers a soft-tissue tumor mass during the freeze cy-cles (12) is a potential benefit that is notavailable with heat-based treatments,which may require the use of special tem-perature-sensitive magnetic resonanceimaging sequences that are performed byusing limited-access open-bore units. CTimaging during freezing is crucial andcan be performed with the straight-shafted cryoprobes that were used in thisstudy, but it requires cranial angulationto avoid the gantry (ie, cranial probe an-gulation also helps to place cryoprobes inupper abdomen locations for percutane-ous hepatic and renal approaches). Ingeneral, the greater the number of cryo-probes used, the larger the lethal ice zone

Figure 5. Transverse CT image findings of cryotherapy for treatment of pulmonary mass andassociated anterior mediastinal adenopathy in a patient in whom chemotherapy was unsuccess-ful. (a) Images show needle placement in left upper lobe nodule (curved arrow) and final extentof ice coverage (straight arrows) in uppermost region of an anterior mediastinal mass abutting theascending aorta (A). No complications were noted, but care was taken to avoid recurrent laryngealnerve damage near the aortopulmonary window—that is, the freezing of this mass did not extendto the course of the nerve wrapping around the left main bronchus. (b) Left: Image shows smallpretracheal nodes (curved arrows) at levels similar to those depicted in a on the right. Straightarrows mark ice coverage. Right: Image obtained at lower level shows ice extension (straightarrows) toward the aortopulmonary window but incomplete tumor coverage (curved arrow)where the posterior aspect of the tumor abuts the pulmonary artery (PA). The tumor was refrozenfollowing regrowth, and local control was achieved for more than 1 year.


Ra

dio

logy

and the shorter the freeze time requiredto achieve thorough tumor coverage bylow-attenuating ice. The suggested freezetimes given in Figure 6 enable one to takeadvantage of the freezing capacity of thenewer 2.4-mm-diameter L-shaped cryo-probes, which allow device insertion andprocedural monitoring with CT fluoros-copy or spiral CT after each freeze, thaw,or refreeze cycle. The ability to place ad-ditional cryoprobes during the thaw cy-cle to cover any margins suspected of be-ing undertreated also helps to alleviateany freezing variations related to tumormargins near major blood vessels or anyfreezing variations along an irregular outermargin. For these situations, we advocatethe eccentric placement of cryoprobescloser to these vessels to intensify the freez-ing and avoid perfusion-mediated heating,without fear of vessel damage because thecollagenous wall architecture of the vesselsis preserved. However, great care shouldstill be taken to avoid any vessel punctureby choosing the least vascular path dis-cerned with lung window imaging set-tings.

Visualization of the treated area be-tween freeze cycles allows additionalcryoprobes to be placed during the thawphase so that one can thoroughly cover atumor during one session rather thanblindly hope that empiric estimates of icesizes in different tissues will still includeall tumor margins. CT fluoroscopy dur-ing PCT now enables rapid and accurateplacement of multiple cryoprobes, aswell as intermittent monitoring of freezeextent. However, careful assessment ofthe freeze extent during the procedure isbest achieved by performing limited spi-ral CT at the end of both freeze cycles(Figs 3a, 5b, 6) because use of the lowerx-ray beam milliampere setting in the CTfluoroscopy mode reduces spatial resolu-tion and tissue contrast (ie, greater quan-tum mottle). Another immediate modifi-cation involves the use of the stick mode(at approximately �10°C) setting on thecryotherapy unit during the thaw cycleto prevent any cryoprobe movement—and associated inaccuracies—betweenfreezes.

An area of cavitation larger than anoriginal peripheral tumor indicates thor-ough ablation coverage (2,12,15) andshould not be mistaken for an abscess,which would have air and fluid levels andthicker, enhancing rims. None of thestudy patients had clinical signs or symp-toms of abscess or sepsis, and 80% (77 of96) of lesions resolved asymptomaticallywithin a month. An area of cavitationlarger than the original tumor is a good

outcome that emphasizes the need forapproximately 1-cm ablative margins astreatment goals. Cavitation surroundingall previous tumor margins also mayserve as an indicator of thorough treat-ment effect, and in our study, 57% (35 of61) of the treated areas were reduced bymore than 50% at 1 month.

The CT appearances of the treated ar-eas over time suggested that even in casesof visible undertreatment, gross tumorregrowth may not be seen until after 3months. This observation is very similarto those described in reports of radiofre-quency treatment: irregular enhance-ment requiring re-treatment within 3–6months (16). Owing to the small samplesize of cases with available follow-updata, it is currently premature to assessany differential responses between tumortypes or to assess the findings in patientswho potentially had tumor regrowth af-ter incomplete treatment. The matchingof the outcomes summarized in Table 2with the observations presented in Table3 will be addressed in future efficacy stud-

ies. Detailed positron emission tomogra-phy data in conjunction with thin-sec-tion contrast material–enhanced CT datamay be required to define the short-termimaging parameters that can be used tobetter predict long-term clinical out-comes.

The 12% pneumothorax rate for PCTsuggests the self-limited nature of cryother-apy-induced pneumothoraces, nearly halfof which were simply evacuated during theprocedure. Cryotherapy can preserve thecollagenous architecture of virtually anyfrozen tissue (3–8), and, thus, it may facil-itate the rapid natural closing of cryoprobetracks. Data are limited for both ends of thetemperature-related treatment spectrum,but radiofrequency cryoprobes could cau-terize tissue channels and become moreprone to causing persistent—or larger—pneumothoraces that progress to broncho-pleural fistulas (11). Animal studies may bebeneficial for assessing these possibilities,but again, such studies may be impracticalbecause the tumor masses in the animalsmay not be of sufficient size for relevant

TABLE 2Tumor Location, Size, and Cryotherapy Coverage

Tumor Location and Diameter (cm) No. of Masses Percentage Tumor Area Covered by Ice*

Peripheral�4 101 98.7 � 0.8�4 65 87.2 � 1.6

Central�4 10 89.5 � 3.8�4 58 79.6 � 2.3

* Relative area ratios of low-attenuating ice compared with soft tumor tissue, based on bidimen-sional measurements and separated according to tumor size and location. Values are expressed aspercentages � standard deviations. They reflect the average cross-sectional area of a tumorcovered by low-attenuating ice per group and indicate a highly significant decrease in tumorcoverage by ice with increasing tumor size and/or central location. Undertreatment was more likelyfor larger central tumors. P � .001 for difference in values between masses 4 cm or smaller andthose larger than 4 cm and for difference in values between the central and peripheral tumors.

TABLE 3Tumor Cavitation and Treatment Size Over Time

Feature 1 Week 1 Month 3 Months 6 Months 12 Months

CT scans available 96 61 76 56 12Tumor appearance

Cavitation 77 (80) 16 (26) 5 (7) 1 (2) 0 (0)Enlargement 87 (91) 12 (20) 4 (5) 6 (11) 2 (17)Size unchanged 8 (8) 14 (23) 18 (24) 23 (41) 4 (33)�50% Involution 1 (1) 35 (57) 47 (62) 25 (45) 4 (33)Disappeared 0 (0) 0 (0) 7 (9) 2 (4) 2 (17)

Note.—Data are numbers of treated masses. Numbers in parentheses are percentages based on thenumber of masses for which there were available scans at the given time point. The number ofmasses for which there were available scans became limited over time, but the images showed cleartrends in terms of two separate imaging features over time: cavitation and size of treatment area.Cavitation nearly resolved by 3 months. Note that cavitation time course correlated well withtreatment site enlargement, reflecting initial tissue destruction beyond the tumor and then invo-lution or stability of the treatment site by 6 months in 86% of cases.


Ra

dio

logy

comparisons with tumor masses in hu-mans.

The other side effects and complica-tions of PCT seen in this study reflect

their transient nature and the associatedvery low rate of progression to seriousconditions that require treatment. Minorhemoptysis resolved within 1 week anddid not require further intervention. Thehypertension that occurred during activefreezing, which accounted for 33% of thecomplications, was mild to moderate andself limited; however, the occurrence ofthis complication should be noted in pa-tients who have preexisting hyperten-sion and thus may require additionaltreatment adjustments. The nerve dam-age seen in two patients reinforces theneed for careful planning of the treat-ment near the brachial plexus and sup-ports concerns regarding recurrent laryn-geal nerve damage when the freezinginvolves the aortopulmonary window.The return of speech in this patient isconsistent with previous reports of po-tential nerve regeneration due to thesheath remaining intact (17), unlike theintegrity of this structure following heat-based ablation or surgery. The two deathsappeared to be temporally related andsuggest the need for caution and carefulfollow-up of patients with limited pul-monary reserve.

Cryotherapy can preserve the collage-nous architecture of the central bronchiand the vasculature, as suggested by re-ported endobronchial cryotherapy expe-riences (3–8). The ability to perform tu-mor ablations immediately adjacent tocentral bronchi may represent a majordifference between cryotherapy andheat-based treatments. New treatmentopportunities for combined mediastinaltherapies may exist, involving cryother-apy in conjunction with lower radiationdoses (18–20) and/or chemotherapy(21,22). Thus, future treatments for ad-vanced-stage lung cancers may havelower overall side effects (eg, minimalscarring) owing to the use of debulkingcombined with cryotherapy. In addition,the patients with advanced disease in thisstudy felt better—that is, their KPS scoresincreased—within 1 week after undergoingcryotherapy. We are uncertain whetherthis outcome was simply a placebo effect orwas related to tumor debulking and a re-sultant reduction in paraneoplastic effects.Thorough staging (ie, not just CT staging),improved ablation planning and monitor-ing, longer follow-up periods (eg, with CT,PET, and extended KPS scoring), and po-tential treatment combinations need to beaddressed for future cryotherapy protocols.

The main limitation of this report isthe lack of long-term follow-up data onthe study patients, who predominantlywere ineligible for surgery and had un-

TABLE 4Complications of PCT

ComplicationNo. of

Procedures* Comments†

Cough 171 (79) Only five (3%) cases exaggerated after cryotherapyHemoptysis 134 (62) Stopped within 1 week; no intervention (embolization) requiredFever 91 (42) Slight or moderate (�38.5°C); resolved within 1–5 days with

antiinflammatory medicationHypertension 72 (33) Self limited and mild to moderate; occurred only during lesion

freezingPleural effusion 30 (14) More frequent with pleura-based tumors; five (17%) effusions

drainedPneumothorax 26 (12) Twelve (46%) cases required periprocedural evacuation, 11

(42%) cases clinically unimportant, three (12%) casesrequired catheter drainage for 5–7 days

Subcutaneousemphysema

11 (5) Commonly occurred in elderly patients; absorbed in 3–5 days

Skin injury 10 (5) Minimal at puncture site; easily prevented by using water-filledglove

Death 2 (1) Neither intraprocedural; one case owing to ARDS and one caseowing to pulmonary embolus

Arm paralysis 1 (.5) Brachial plexus damage from direct approachLoss of speech 1 (.5) Temporary aphasia from recurrent laryngeal nerve damageAll 217 (100) . . .

Note.—The more common complications were self-limited and did not require intervention orsymptom treatment, whereas the more severe complications (eg, pneumothorax) were relativelyuncommon given the extent and complexity of PCT in this series.

* Numbers in parentheses are percentages.† ARDS � acute respiratory distress syndrome.

Figure 6. Illustration of current PCT equipment and outline of pro-tocols for thorough freezing. The number of cryoprobes and theprojected freeze schedules relate to tumor diameter. The dashed lineinside the ice balls indicates increasing extent of lethal (eg, �20°C) icecoverage because more cryoprobes are used to cover the mass. Thebracketing approach (ie, with use of two cryoprobes) for treatment ofsmall (�3 cm) masses helps to eliminate operator dependency in theprecise placement of cryoprobes in the center of the tumor and toextend cytotoxic margins, and it may facilitate shorter freezing times.Projected ice ball sizes vary according to the tumor’s location nearmajor vessels and the background tissue in which it resides (eg, lung,liver, kidney, or bone). Diam. � diameter, Tx � treatment.


Ra

dio

logy

dergone other treatments that failed.Therefore, this was not a definitive trialfor assessing the effectiveness of PCT fortreatment of either newly diagnosed pri-mary cancers or metastatic lung tumors.Our report of the large number of proce-dures for which imaging findings wereavailable should serve only as a thoroughdocumentation of safety, feasibility, andtechnique considerations. Further PCTexperiences with fluoroscopic CT guid-ance and monitoring performed by usingthe newer CT-compatible cryoprobes areneeded to minimize or eliminate visiblyundertreated areas of larger and centraltumors, such as those seen in this study.We also suggest that interventionists en-gage in more standardized cryotherapyplanning to better adapt to the continu-ally improving cryotechnology and tomitigate operator-dependent outcomes.Despite some early technical shortcom-ings, the feasibility of achieving cryo-therapeutic control of local tumor growthwas safely demonstrated in the describedlarge group of patients, who could not un-dergo surgery and/or had previously un-dergone frontline therapy that failed.

In summary, CT-guided PCT was asso-ciated with low procedural morbidity,even with freezing near mediastinalstructures, which also appear to healwithout substantial scarring or sequelae.Ongoing advances in cryotechnologyand imaging guidance techniques mayhelp interventionists avoid the degree ofundertreatment of larger central massesobserved in the current study and im-prove treatment planning.

References1. Bahn DK, Lee F, Badalament R, Kumar A,

Greski J, Chernick M. Targeted cryoabla-

tion of the prostate: 7-year outcomes inthe primary treatment of prostate cancer.Urology 2002; 60(2 suppl 1):3–11.

2. Lee FT, Mahvi DM, Chosy SG, et al. He-patic cryosurgery with intraoperative USguidance. Radiology 1997; 202:624–632.

3. Sanderson DR, Neel HB, Fontana RS.Bronchoscopic cryotherapy. Ann OtolRhinol Laryngol 1981; 90:354–358.

4. Homasson JP. Bronchoscopic cryother-apy. J Bronchol 1995; 2:145–153.

5. Maiwand MO. The role of cryosurgery inpalliation of tracheo-bronchial carci-noma. Eur J Cardiothorac Surg 1999; 15:764–768.

6. Mathur PN, Wolf KM, Busk MF, Briete M,Datzman M. Fiberoptic bronchoscopiccryotherapy in the management of tra-cheobronchial obstruction. Chest 1996;110:718–723.

7. Deygas N, Froudarakis, Ozenne G,Vergnon JM. Cryotherapy in early super-ficial bronchogenic carcinoma. Chest2001; 120:26–31.

8. Littrup PJ, Mody A, Sparschu RA, et al.Prostatic cryotherapy: ultrasonographicand pathologic correlation in the caninemodel. Urology 1994; 44:175–184.

9. Dupuy DE, Zagoria RJ, Akerley W, Mayo-Smith WW, Kavanagh PV, Safran H. Per-cutaneous radiofrequency ablation of ma-lignancies in the lung. AJR Am JRoentgenol 2000; 174:57–59.

10. Zagoria RJ, Chen MY, Kavanagh PV, TortiFM. Radio frequency ablation of lung me-tastases from renal cell carcinoma. J Urol2001; 166:1827–1828.

11. Dupuy DE, Mayo-Smith WW, Abbott GF,DiPetrillo T. Clinical applications of ra-dio-frequency tumor ablation in the tho-rax. RadioGraphics 2002; 22:S259–S269.

12. Lee FT Jr, Chosy SG, Littrup PJ, WarnerTF, Kuhlman JE, Mahvi DM. CT-moni-tored percutaneous cryoablation in a pigliver model: pilot study. Radiology 1999;211:687–692.

13. Liu BQ, Peto R, Chen ZM, et al. Emergingtobacco hazards in China. I. Retrospec-tive proportional mortality study of onemillion deaths. BMJ 1998; 317:1411–1422.

14. Littrup PJ, Wang H, Duan Y, Zhang Y,Feng H, Nie Z. Percutaneous cryotherapy

of the thorax: clinical observations frommore than 200 procedures (abstr). Radiol-ogy 2003; 229(P):438.

15. Rodgers BM, Blake KD, Alexander JA. Theeffects of profound cryotherapy upon thepulmonary parenchyma. J Thorac Car-diovasc Surg 1982; 83:784–789.

16. Gervais DA, McGovern FJ, Arellano RS,McDougal WS, Mueller PR. Renal cell car-cinoma: clinical experience and technicalsuccess with radio-frequency ablation of42 tumors. Radiology 2003; 226:417–424.

17. Moorjani N, Zhao F, Tian Y, Liang C, Ka-luba J, Maiwand MO. Effects of cryoanal-gesia on post-thoracotomy pain and onthe structure of intercostal nerves: a hu-man prospective randomized trial and ahistological study. Eur J CardiothoracSurg 2001; 20:502–507.

18. Fukumoto S, Shirato H, Shimzu S, et al.Small-volume image-guided radiotherapyusing hypofractionated, coplanar, andnoncoplanar multiple fields for patientswith inoperable stage I nonsmall celllung carcinomas. Cancer 2002; 95:1546–1553.

19. Vergnon JM, Schmitt T, Alamartine E,Barthelemy JC, Fournel P, Emonot A. Ini-tial combined cryotherapy and irradia-tion for unresectable non-small cell lungcancer: preliminary results. Chest 1992;102:1436–1440.

20. Chen A, Galloway M, Landreneau R, et al.Intraoperative 125I brachytherapy forhigh-risk stage I non-small cell lung car-cinoma. Int J Radiat Oncol Biol Phys1999; 44:1057–1063.

21. Homasson JP, Pecking A, Roden S, Ange-bault M, Bonniot JP. Tumor fixation ofbleomycin labeled with 57 cobalt beforeand after cryotherapy of bronchial carci-noma. Cryobiology 1992; 29:543–548.

22. Clarke DM, Baust JM, Van Buskirk RG,Baust JG. Chemo-cryo combination ther-apy: an adjunctive model for the treat-ment of prostate cancer. Cryobiology2001; 42:274–285.


Ra

dio

logy

thoracic masses treated with adiology percutaneous ... · endobronchial cryotherapy has been...

Documents