124: Photodynamic Therapy in the Management of Pleural Tumors

Cancer of the pleura is a virulent and lethal malignancy. Primary tumors of the pleura are rare, whereas metastatic tumors, often in the form of malignant pleural effusions, are quite common. Primary tumors of the pleura, malignant pleural mesothelioma being the most common, are typically associated with a life expectancy of less than a year. Metastatic cancers to the pleura, including non–small-cell lung cancer (NSCLC), represent stage IV disease and usually coincide with the most adverse prognosis for the primary cancer that has spread to the pleura.

The most common form of treatment for pleural cancers is systemic therapy and/or palliative care, since the majority of pleural cancers represent metastatic disease. Surgery is not typically considered an effective treatment because of the essentially impossible task of achieving a true negative margin for these cancers that coat every surface of the chest cavity. In an investigational capacity, however, surgery has become the cornerstone of highly aggressive multimodal treatment plans in selected patients, and the most widespread application in malignant pleural mesothelioma.

With the expectation that microscopic disease will remain after even the most aggressive surgical resections, one approach has been to combine an intraoperative adjuvant therapy with systemic therapy and, sometimes, adjuvant external beam radiation. The intraoperative adjuvant therapies include the following: chemotherapy, with or without hyperthermia, heated povidone iodine, radioisotopic radiation, intraoperative photon radiation, and photodynamic therapy (PDT).¹ This chapter focuses on the combination of surgery and PDT and the application of this technique in malignant pleural mesothelioma.

PDT is a technique for killing tumors which uses a photosensitizer that is activated by visible light. It has been observed that photosensitizers are preferentially taken up by, or retained in, tumor cells.^2,3 Once inside the cells, the photosensitizer is activated by a laser light with a wavelength specific to the sensitizer's absorption spectrum. Activation of the photosensitizer in the presence of oxygen results in the production of excited species of oxygen capable of inducing cell death. Cell death occurs by apoptosis or from direct destruction of certain cellular elements.^4,5 In addition, PDT may result in neovascular damage that may compromise the tumor's blood supply.⁶ Finally, when PDT is used to treat cancer, it appears to enhance the host's immune response to the tumor.^7,8

In addition to the presence of oxygen, the items needed to perform pleural PDT include the photosensitizer, a light source, and a dosimetry system. PDT is a dose-dependent treatment. That is, without light activation there is no effect on cells that contain the photosensitizer. The overall effect of PDT increases with the amount of activating light delivered. Although photosensitizers may demonstrate some selectivity for neoplastic cells, they also partition into normal tissues and will cause some degree of damage to those tissues when exposed to light. This is critical, especially for pleural PDT, as injury to any structure in the chest can occur absent meticulous attention to light dosimetry.

Photosensitizers

The majority of experience with pleural PDT has been with two photosensitizers, Photofrin and Foscan. Photofrin (dihematoporphyrin derivate) was the first commercially available photosensitizer and has its major excitation wavelengths in the UV region (200–450 nm), the green region (510 nm), and a small absorption peak in the red (630 nm) region of the light spectrum. Meta-tetrahydroxyphenylchloride (Foscan®) is the second drug that has been used for treating mesothelioma.² It has major absorption peaks in the UV (200–450 nm) green regions (520 nm), with the highest peak in the red region at 652 nm. Although the shorter wavelengths have higher energy, red light is used for greater tissue penetration, which can be up to several centimeters depending on the absorption and scattering characteristics of the tissue.

Both of these photosensitizers are administered intravenously, preoperatively. As stated previously, photosensitizers may demonstrate some element of selectivity, but will ultimately sensitize all tissues, including the skin. As a result, cutaneous photosensitivity is the primary toxicity associated with administering a photosensitizer.^9,10

Laser Equipment

To treat large surface areas with PDT, high-power light sources are required. In general, it is necessary to use a laser to supply light at the appropriate wavelength and intensity. Tunable dye lasers, pumped by a larger green light laser with fixed wavelength, are commonly used to produce red light in the 7 W range. These lasers have the advantage of possessing dye modules that can be interchanged to permit production of a broad spectrum of wavelengths. The disadvantage is that they are relatively large and require high-power supplies and water-cooling systems. Recently developed diode lasers are more portable and have power outputs up to 6 W (in the red light waveband). They do not require the use of high-power supplies or water-cooling systems, but have the disadvantage of being fixed at a single wavelength.

Dosimetry

Although photosensitizers may preferentially migrate to tumor tissues, it should be assumed that all tissues are photosensitive. As a result, any structure that is illuminated can be injured. It is crucial, therefore, not to overdose normal tissues with light. Some investigators rely upon “calculated” light doses.^11,12 On the basis of experiments which demonstrate that the measured and calculated light doses may vary widely owing to the unpredictable reflection and refraction patterns of light in vivo,¹³ we believe that light dosing should not be empiric and one should rely on measured dosimetry. Consequently, light sensors are placed at strategic positions within the hemithorax and fed into a real-time dosimetry system that has a separate channel for each sensor. During PDT, the light source is moved around the chest cavity until each sensor has measured the desired dose of light.

Currently two types of sensors are used: flat and isotropic. The flat sensors¹⁴ underestimate the light dose delivered to the tissue surfaces in comparison with isotropic detectors.¹⁵ If the type of sensor or photosensitizer is changed, measurements must be made to determine the conversion factor between the sensors or safety studies to determine the safe maximal tolerated dose.¹⁶

Some investigators fill the hemithorax with diffuse intralipid solution to help scatter light as the light source is moved around the chest cavity.^14,17 This is our preferred method for light delivery regardless of the debulking technique. It assures that there is no shielding of tissue by pooled blood and also permits direct manipulation of the costophrenic recesses, the most difficult areas in which to achieve good light delivery. Others have focused on integral illumination by using a bulb fiber and no light-diffusing medium.¹³ This technique is only applicable after a pneumonectomy. In the latter technique, a transparent sterile bag is placed in the chest cavity following pneumonectomy and filled with warm saline to facilitate flattening and expansion of the chest cavity structures. After partial closure of the surgical wound, a single spherical bulb fiber is placed in the center of the bag to allow integral illumination of the entire cavity and enhance the reflection of light. This technique is not compatible with a lung-sparing procedure and may not be applicable if it does not appear that the bag will expand all crevices in the hemithorax. We abandoned this technique when we found that blood was pooling under the bag and preventing light from reaching those areas.

Overview

The concept behind the combination of surgery and PDT for pleural malignancies is that surgery is used to achieve a macroscopic complete resection and PDT is performed after the resection as an intraoperative adjuvant therapy in an effort to treat the residual microscopic disease.^18–21 The two options for achieving a macroscopic complete resection include extrapleural pneumonectomy (Chapter 122) and radical pleurectomy (Chapter 121).

Extrapleural pneumonectomy is defined as the en bloc removal of the lung, parietal pleura, diaphragm, and pericardium. Typically, both the diaphragm and pericardium are reconstructed with prosthetic patches. This operation has the advantage of being standardized with respect to both name and technique. It is almost certainly the technique that results in the least amount of residual microscopic disease, that is, the most complete and reproducible macroscopic complete resection. Finally, without the lung in the chest, radiation can be used as an adjuvant therapy to treat the entire hemithorax.

Radical pleurectomy is a more nebulous operation which, even in the best of hands, almost certainly leaves behind more microscopic disease than an extrapleural pneumonectomy. There is essentially no standardization of this operation and, in fact, the procedure appears in the literature under a multitude of names including pleurectomy, decortication, pleurectomy-decortication, radical pleurectomy-decortication, radical pleurectomy and extended pleurectomy. The intent of the operation also varies, from a palliative debulking of some gross disease to a macroscopic complete resection. This tremendous variability in every aspect of the procedure, including nomenclature, makes it difficult to compare published case series. Finally, the timing of the decision to perform the operation is also variable, ranging from an intraoperative decision based upon intraoperative findings to a preoperative plan. In the former situation, either the bulk of the cancer or the degree of involvement of the cancer with the pulmonary fissures is often cited as the deciding factor as in whether or not the lung can be saved.

For the purposes of the ensuing discussion, radical pleurectomy is the term that we use to describe a lung-sparing operation aimed at achieving a macroscopic complete resection. The goal of each procedure is to save the lung and, if possible, the phrenic nerve and as much of the pericardium and/or diaphragmatic musculature as possible, while still achieving a macroscopic complete resection. Depending upon the degree of invasion, it usually is not necessary to reconstruct the diaphragm. We have found this to be true for the pericardium as well, since the presence of the lung is sufficient to prevent cardiac herniation/torsion. In our hands, radical pleurectomy is a procedure that is planned preoperatively, not an intraoperative decision based upon involvement of the pulmonary fissures, tumor bulk, or other factors.

Thus, with respect to the two types of surgery for malignant pleural mesothelioma, the advantages of extrapleural pneumonectomy include relative standardization and hence comparability of results, the best macroscopic complete resection, the ability to treat with adjuvant radiation and, in our hands, a more expeditious operation than radical pleurectomy. The disadvantages are the consequences of pneumonectomy and, potentially, the need for prosthetic reconstructions. The principal advantage of radical pleurectomy is preservation of the lung and, potentially, the decreased need for prosthetic reconstruction. Preserving the lung not only translates into the potential benefits of preserving quality of life and offering a surgery-based approach to patients who might not be candidates for pneumonectomy, but it may also allow the patient to undergo more aggressive treatment options for their inevitable tumor recurrence. The disadvantages of radical pleurectomy include the presence of more residual microscopic disease, lack of standardization and, in our hands, a longer operation with a more complex postoperative management than with extrapleural pneumonectomy. The ideal surgical approach remains an area of controversy. It may well be that no one approach is correct for every patient. The optimal operative strategy may well be related to the individual patient, characteristics of their particular tumor, and the selection of adjuvant therapies that are going to be used in combination with surgery.

We have performed both extrapleural pneumonectomies and radical pleurectomies for mesothelioma in combination with intraoperative PDT. In a pilot study that compared the outcomes of these two surgical approaches, we found that the patients who underwent radical pleurectomy had longer overall survival compared with patients who underwent pneumonectomy (Fig. 124-1A), despite essentially no difference in disease-free survival (Fig. 124-1B). The term “MEPP” was used in the study cited in Figure 124-1 as the operation involving pneumonectomy preserved the diaphragm, pericardium, and phrenic nerve. The currently accepted definition of extrapleural pneumonectomy is en-bloc resection of the parietal and visceral pleura with the ipsilateral lung, pericardium, and diaphragm. In cases where the pericardium and/or diaphragm are not involved by tumor, these structures may be left intact.²²

On the basis of this pilot study, we switched exclusively to radical pleurectomy as our surgical approach to malignant pleural mesothelioma. We still, however, will perform an extrapleural pneumonectomy in combination with PDT for NSCLC with pleural dissemination (stage IVa). Although the era of targeted therapies has provided more treatments for patients with this disease, this approach has shown promise as an aggressive option for patients with this cancer.²³ In either case, whether the lung is taken or spared, the light precautions taken during surgery and the performance of PDT are identical.

Patient selection for these procedures takes place in the forum of a multidisciplinary tumor board-type conference with disease limited to one hemithorax as the principal oncologic criteria for consideration as a surgical candidate. Eligible and interested patients undergo an extensive radiographic staging workup and ultimately undergo an invasive staging procedure including bronchoscopy and laparoscopy to rule out radiographically occult metastases. While the presence of ipsilateral mediastinal lymph node metastases (N2 disease) is not currently viewed as an exclusion criterion for radical pleurectomy for mesothelioma, as it is for extrapleural pneumonectomy, this does correlate with a decrease in overall survival and we have started to routinely include endobronchial ultrasound-guided biopsy (EBUS) staging as part of our preoperative evaluation. Contralateral thoracoscopy, to rule out contralateral pleural disease, and mediastinoscopy or EBUS, to rule out N3 disease, are used on a case-by-case basis as dictated by the imaging studies and clinical suspicion. From a safety perspective, the selection criteria are the same as would be used for any major thoracic operation, such as a formal decortication or pneumonectomy, with an added emphasis on nutritional parameters. The PDT superimposes a significant metabolic demand, and malnutrition serves as an exclusion criteria. As part of the informed consent disclosures, it is made clear to all patients that the procedure is investigational and, in addition to the risks of the surgery, there are the superimposed risks of PDT; primarily, cutaneous photosensitivity and a higher incidence of postoperative atrial fibrillation, deep venous thromboses (but not pulmonary embolism), and persistent air leaks in radical pleurectomy patients.

Photosensitizer and Light Precautions

Before surgery, the patient receives the photosensitizer as an outpatient. The patient becomes immediately light sensitive. Therefore, patients should be instructed to bring sunglasses and wear appropriate clothing to cover or shade all exposed skin. With adequate patient education, we have not experienced any problems with sunburning before or after surgery. Once the patient arrives for surgery, hospital light precautions are initiated. This includes no exposure to sunlight through windows or intense overhead lights (fluorescent lights are fine) and probe rotation or spot-checking pulse oximetry. In the operating room the overhead lights and surgical headlights are passed through yellow filters. Yellow represents the portion of the visible light spectrum where the photosensitizers absorb less light, but these are not turned on until the incision is shielded with towels and all skin is protected from these intense light sources (see Fig. 124-2).

Technique of Radical Pleurectomy

General Approach and Strategy

Over the years we have tried multiple techniques in an attempt to develop a standardized approach to radical pleurectomy. What follows is a description of our current iteration of this procedure. The general strategy that has resulted in the most reproducible results is to mobilize the entire cancer from the hemithorax, such that it is tethered solely to the lung, and then resect the entire visceral pleura, en bloc with the mobilized cancer. With the proviso that every one of these cancers is different and the surgeon must remain flexible in the approach, the typical order of dissection is bony hemithorax, posterior mediastinum, superior mediastinum, anterior mediastinum, diaphragm, and lung.

Preparation and Incision

Light precautions, as stated above, are taken from the time the patient receives the photosensitizer. Patients undergoing radical pleurectomy will need a central line with one port reserved for total parenteral nutrition, epidural catheter, ipsilateral radial and femoral arterial lines and a nasoenteric tube, peripheral venous access, and a Foley catheter. Once it is confirmed by bronchoscopy that the nasoenteric tube is not in the lungs, it is our routine to give 60 mL of heavy cream spiked with an amp of methylene blue to aid the detection of injury to the thoracic duct during the course of the surgery. The patient is then placed in the lateral decubitus position and a thoracotomy incision is created under operating room fluorescent lights only. The latissimus is divided, but we are usually able to preserve and retract the serratus muscle. Once the chest wall layer is approached, the towels can be sewn to the muscle fascia to shade the skin, and the overhead and surgical headlights can be activated. If there is a rib interspace, the extrapleural plane is approached through the sixth interspace. If the interspace is contracted to the point of rib overlap, or the patient has had previous surgery through the sixth interspace to preclude entry, the seventh rib is removed and the extrapleural plane is approached through the bed of the resected seventh rib.

Chest Wall/Posterior-Superior Mediastinal Mobilization

The initial portion of the operation is the same whether the surgeon is planning to perform a radical pleurectomy or extrapleural pneumonectomy. The first step of the mobilization is to free the cancer from the bony hemithorax, followed by the posterior and superior mediastinum. The plane is identified and entered adjacent to the incision. It is developed bluntly, as much as possible. Blunt finger dissection, working a broad front, causes cleavage in the correct plane. Sharp dissection is more likely to leave behind gross tumor. The argon beam coagulator or Aquamantys® is good for cauterizing the chest wall, from which capillary oozing can lead to significant blood loss. Dissecting the chest wall is the safest portion of the operation and provides a good opportunity for the surgeon to get a sense of how the tumor is interacting with the tissues (Fig. 124-3A,B).

As the dissection is carried to the posterior reflection onto the posterior mediastinum, the surgeon can follow the intercostal veins from the azygos or hemiazygos veins as they traverse the mediastinum to safely transition from the chest wall to the posterior mediastinum. On the right side, the surgeon must take care not to get behind the esophagus; on the left it is the aorta that must be left in place as the pleura is separated from its medial surface. The presence of the nasoenteric tube can aid in identifying the esophagus by palpation, sometimes earlier in the setting of bulky tumors than can be accomplished by vision. In the superior mediastinum, the surgeon must be mindful of the subclavian artery on the left and the vena cava on the right. A 30-degree video thoracoscope for supplemental vision during the dissection is often helpful to assure the correct plane is identified and maintained, especially in the apex of the chest as the thoracic inlet and most superior mediastinal structures are dissected. On the right side, the azygocaval junction is typically approached both superiorly and posteriorly, having followed the azygos vein from behind and the cava from above. A venous injury can occur in this area if the wrong plane is entered.

Anterior Mediastinum

The anterior mediastinum is approached by sweeping off all pericardial fat in an anteroposterior direction, starting in the pericardiosternal recess. The surgeon must take care not to breach the anterior pleural reflection and enter the opposite hemithorax. This portion of the operation is highly variable. Occasionally, nearly the entire pericardium is covered with pericardial fat and removing this fat leaves the pericardium with a macroscopic complete resection. More commonly, the pericardial fat dissection gets the surgeon out of the anterior recess, but most of the pericardium is found to be directly involved with the cancer. Rarely, the tumor will separate, leaving normal appearing pericardium. If that is not the case, an attempt can be made to separate the layers of the pericardium, leaving the serous pericardium intact and resecting the fibrous pericardium en bloc with the cancer and mediastinal pleura. While technically challenging, it is often possible. If the pericardium remains intact, it should be fenestrated after the PDT to avoid postoperative tamponade. If the pericardium is too extensively involved to achieve a macroscopic complete resection, the surgeon has two options: a small area can be resected and the surgeon may wish to sew in a prosthetic pericardial patch if there is a concern about potential cardiac herniation, or, for an extensive area, the entire pericardium can be resected. As the lung is being left in place, the pericardium does not have to be reconstructed unless the surgeon is concerned that the patient will be at risk for cardiac torsion or herniation or if it is their preference. Regardless of approach, the goal remains to achieve a macroscopic complete resection. In either case, if there is an area of full thickness involvement of the pericardium that will require resection, that area is left on the pericardium until after the PDT is administered to avoid directly illuminating the heart. After concluding the PDT, the pericardium can be removed with the same options depending on the amount resected as discussed previously. If the surgeon wishes to place a patch, this is done in the same fashion, as with a formal extrapleural pneumonectomy, and it should be pie-crusted or fenestrated to avoid tamponade.

As the dissection is extended posteriorly, ultimately culminating at the anterior hilum of the lung, the phrenic nerve can be identified at the level of the superior mediastinal dissection and skeletonized as it traverses the anterior mediastinum. Whether or not preserving the nerve and the diaphragm without a pleura helps with respiratory function postoperatively is a current area of investigation. It is intuitively attractive and remains our current approach to preserve the phrenic nerve, which is usually possible, even when encased in bulk tumor (Fig. 124-4).

Diaphragm

The diaphragm dissection is started in the costophrenic recess, attempting to bluntly separate the pleura from the underlying bare musculature. Occasionally, rarely, the pleura will separate from the underlying bare musculature. Often, however, this requires sharp dissection and is best accomplished with broad-tipped scissors, allowing the scissors to “find” the plane between the hard cancer and the soft underlying normal tissue (see Fig. 124-3B).

Limited areas of full thickness invasion are resected, leaving only peritoneum, and the diaphragm can be reconstructed primarily with heavy absorbable sutures. Sometimes, especially if the inseparable tumor is a central island and there is sufficient laxity in the remaining debrided diaphragm, the area can be tented away from the abdomen and undercut with a thick tissue stapler. Care must be taken to assure that no viscera are caught in the stapler. This is readily done by palpation. The staple line is then oversewn with heavy absorbable sutures as the diaphragmatic muscle, comprising the staple line, can tear and result in a hernia. If too much diaphragm is involved, it must be resected, as with an extrapleural pneumonectomy, and reconstructed with a 2-mm Gore-Tex (W.L. Gore Associates, Inc., Flagstaff, Arizona) patch.

Lung

At this point, in the dissection, the entire tumor is tethered solely to the lung. The anesthesiologist is asked to connect the operative lung to an alternate oxygen supply with an in-line stack of PEEP valves that will allow it to be held under positive pressure ranging from 10 to 30 cm water. The tumor is sharply incised, extending through the visceral pleura (Fig. 124-5A). The plane between the undersurface of the visceral pleura and bare lung parenchyma is developed. Initially, this is best accomplished with a forceps and fine scissors, until at least several millimeters beyond the incision has been liberated. At this point the edge can be better grasped and the bare lung parenchyma can be very gently retracted. The denuded lung is best retracted with a finger and coarse mesh gauze. A well-intentioned assistant can easily plunge a suction catheter into the parenchyma with only minimal pressure and thus should be cautioned, as the lung is devoid of pleura and is astoundingly delicate without the visceral pleura. If this occurs, it will result in air leaks that do not seal very readily. Over the years, the instrument that has proved best suited for further developing the plane is a broad Cobb dissector. Initially, there is often a torrential air leak when the visceral pleura is removed from the parenchyma but, literally, within minutes the leaks nearly abate if the plane has been maintained at the interface of the visceral pleura and parenchyma. In the setting of nodular, rather than planar, cancer the operator simply follows the contour of the tumor as it extends into the lung parenchyma. Often these divots will overlie areas of compressed parenchyma that re-expand when the cancer is separated. Sometimes a cancer will be encountered where portions, or even the entire lung, will not yield the subpleural plane. In these cases, the electrocautery can be used to open the plane that is visibly evident, but does not yield to cold steel dissection. There are rare cases, often in the setting of mixed histology, when a significant portion of the dissection must be performed with cautery.

A critical element in the lung dissection, and what is often considered a contraindication to lung-sparing surgery, is the requirement to remove tumor from the fissures. This is safely accomplished by saving this part of the lung dissection for last. The cancer is tracked down into the fissure and followed from both sides. Surprisingly, the planes within the fissures are often better preserved and more readily separated than on the other surfaces of the lung. An even level along both sides is maintained as the base of the fissure is approached. At this point the lung is deflated and the surgeon will be able to palpate the deepest portion of the fissure, where the cancer has formed a cast. In any patient with complete fissures, this will typically terminate on the surface of the ongoing pulmonary artery. Under direct vision, the veil-like investing tissue over the artery is sharply divided, thereby releasing the cancer from the fissure. This will often result in skeletonization of the pulmonary artery within the fissure (Fig. 124-6).

Once the tumor is released in the fissure, the similar investing band of extrapleural tissue can be identified and divided around the hilum, thereby dividing the remaining attachments of the cancer to the patient. In cases where the cancer is pliable, a single large specimen may result. Sometimes, however, the cancer may be so firm that in order to have enough room to work and still preserve the lung, the cancer is better removed piecemeal.

Once the specimen is removed, the chest cavity should be inspected to assure that a macroscopic complete resection has been achieved. Again, video thoracoscope is very useful to help inspect all surfaces and also provides magnification.

Lymphadenectomy

A thoracic lymphadenectomy is then performed dissecting all standard “numbered” nodal stations as well as any phrenic and/or internal mammary nodes. In addition, the author (JF) has been harvesting the posterior intercostal lymph nodes. The significance of these lymph nodes is not established and is an area of active investigation. For the purpose of our publications and analyses, given that these lymph nodes are not described in any current staging schema, we have considered them N1 lymph nodes as we have had multiple instances where these were the only positive lymph nodes. We do, however, have preliminary data suggesting these posterior intercostal nodes may correlate inversely with survival and therefore encourage others to collect these nodes such that their significance can be established. To access them, use electrocautery to incise the posterior interspaces at the level of the rib heads. Often the nodes can be bluntly delivered with a fingertip, like ejecting a pea from a pod, but sometimes it is necessary to reach into the interspace with a Singley forceps. Care must be taken during this maneuver to avoid avulsing the intercostal vessels with which they are associated.

With the lymphadenectomy completed, a final check is done to confirm hemostasis. At this point, preparations are made to start PDT.

Intraoperative PDT

PDT is a dose-dependent treatment. Without any light, the photosensitizer has no effect. With too much light, injury can occur. The key is to precisely measure the amount of light being delivered. By doing this, we are able to routinely deliver intraoperative hemithorax PDT without any direct PDT injuries. There are multiple complications related to the PDT effect, primarily from cytokine release, but we do not experience “burn” injuries that were encountered in the early experience of PDT or are still occasionally witnessed with Phase I dose escalation studies with new photosensitizers. Because the photosensitizer will absorb photons that are directly incident, refracted or reflected, in a complex and moving geographic space like the chest cavity it is impossible to estimate or calculate the light dosage based on time, distance, and power output of the laser. It is by necessity that the amount of delivered light is measured in real time. This is accomplished with light detectors and a dosimetry system.

The first step is to sew in the light detectors. These are isotropic detectors comprised of a small titanium sphere attached to a thin glass fiber (Fig. 124-7A). These detectors are fed into a length of saline-filled sterile intravenous tubing, which is then folded over on the tip and tied with a silk suture. This provides a ready loop for sewing the tubing into the chest cavity (Fig. 124-7B). Typically seven detectors are placed at strategic locations within the hemithorax: anterior and posterior diaphragmatic sulci, apex, anterior and posterior chest wall, pericardium, and posterior mediastinum. These detectors are brought out from the chest cavity (Fig. 124-7C) and connected to the dosimetry system (Fig. 124-7D). This system is composed of the processing box which collects the data and subsequently feeds that information into a laptop computer where the screen gives a real-time read out of both the current fluence rate being seen by each detector and the total cumulative dose measured at each detector (Fig. 124-7E).

Light delivery is accomplished by using an optical fiber, with the tip terminating in a modified endotracheal tube. The tube and the balloon are filled with dilute (10%) intralipid. This accomplishes several objectives. First, it allows the fragile fiber to be easily maneuvered about the chest cavity. Second, it protects the tissues from the actual fiber tip, which is very hot even though it is just visible light that is being delivered. Third, the balloon, in which the tip is centered, helps disperse the light 360 degrees. Once the light detectors are in place and have been tested, PDT can commence. The procedure is quite simple.

Light of the appropriate visible wavelength (porfimer sodium, for instance, absorbs red light at 630 nm) is shined into the chest cavity. Dilute (0.1%) warm intralipid is poured into the chest cavity. This also accomplishes several objectives. First, the fat globules in the intralipid act to reflect the light, which helps to deliver it into all recesses. Second, the intralipid prevents any pooling of blood, which would absorb the light and hence shield the underlying tissue. Third, the intralipid floats the collapsed lung, making it easier to move the tip of the light detector around all surfaces, especially into the fissures of the lung and the hilar regions. The intralipid is constantly turned over during the light delivery, as it becomes blood tinged. Blood absorbs the light and decreases the efficiency of the light delivery. The light is then simply moved around the chest cavity until each of the detectors has recorded the desired dose of light (Fig. 124-7E). A pitfall is not attempting to get all the detectors to register similar dosimetry values at all times. The amount of time required for light delivery depends on the photosensitizer, but is typically an hour or less for an average-sized patient (Fig. 124-8).

Once the PDT is completed, gowns, gloves, and the shielding towels are replaced using a “clean-contaminated” type strategy. Used instruments are not replaced, but are soaked or wiped down with sterile water during the PDT. If a macroscopic complete resection was achieved prior to the PDT, then closure can commence. If a portion of full thickness tumor was left on the diaphragm or pericardium, to avoid directly illuminating the heart or abdominal viscera, it is resected at this point. If the defect in the diaphragm is sufficiently large, a patch can be placed. Most often, however, there is enough laxity in the diaphragm to permit primary closure. If the pericardium was completely debrided, without disruption, it is fenestrated to avoid potential tamponade. If any part of the pericardium requires resection, the tenets and options discussed earlier are followed. Finally, if no chyle leak was detected during the course of the operation, as indicated by the administered cream/methylene blue, then no action need be taken with the thoracic duct. If on the right side the surgeon is concerned about duct disruption, despite the lack of chyle in the field, then the duct can be presumptively ligated at its emergence through the aortic hiatus.

Straight chest tubes are placed anteriorly and posteriorly to the apex. Additional holes should be cut to access air and fluid throughout the entire intrathoracic traverse of the tubes. A rongeur is helpful for making extra holes in the chest tubes, taking care not to cut more than a quarter the diameter of the tube to avoid kinking, and making sure the most proximal hole is cut through the radiopaque marker to assure all holes are intrathoracic. A right angle tube or flexible fluted tube is also placed along the diaphragm, terminating in the posterior costophrenic recess.

Depending upon standard criteria, as well as the surgeon's sense of the need for positive pressure to maintain full lung expansion, the patient can either be extubated or remain on pressure mode ventilation at the end of the operation. There are usually a lot of bloody secretions at the conclusion of these procedures, so the patient should undergo a complete toilet bronchoscopy at the conclusion of the operation and, again prior to extubation if the patient remains intubated. If the patient is extubated, suction should be placed at -20 cm on each chest tube, and even increased if the lung is not fully inflated on the postoperative chest x-ray. There is a premium on achieving full lung expansion and the higher suction to achieve full lung expansion can, counterintuitively, help the air leaks seal more quickly. If the patient is left intubated, usually -10 cm of suction is adequate and can be increased when the patient is extubated. Because of the air leaks, volume calculations on the ventilator are unreliable and the ventilator needs to be adjusted empirically, based upon blood gasses. Typically, patients are hypocarbic, likely because of the air leaks. Having all three chest tubes connected to separate collection devices will allow the surgeon to assess leaks and drainage in a more useful manner. If the tubes are still leaking after 2 days, the suction can be decreased as long as the lung remains fully inflated on chest radiograph. Once on water seal, the leaks tend to stop quickly. Tubes are then removed per routine criteria. Despite the enormity of the initial leaks, persistent air leaks occur less than 10% of the time, and we have sent patients home with Heimlich valves, but have never had to reoperate to address a persistent leak.

Attention to nutrition and pulmonary toilet are the critical issues postoperatively. Because most patients will have a compromised diaphragm, even the most motivated patients may have trouble clearing secretions and may require one or several awake bronchoscopies.

Light precautions, depending upon the photosensitizer, must be continued. Perhaps the single most critical postoperative PDT-related consideration is that extreme caution must be taken surrounding the esophagus. PDT causes the esophagus to become very fragile, but it can recover fully and without incident if there is no manipulation. Early in our experience, we had two esophageal perforations that resulted in mortalities. The first was simply from the replacement of a nasogastric tube. The second was from an upper endoscopy performed for upper GI bleeding. Consequently, for the first 6 weeks we are extremely cautious and try to avoid any esophageal manipulation, even replacing a nasogastric tube. With this policy, we have had no additional esophageal complications. Additional postoperative considerations surrounding the addition of PDT to this already sizable operation are primarily related to the inflammatory response and additional metabolic demand. Hence, even patients with normal preoperative albumen will typically drop to half normal levels making the attention to nutrition, starting with immediate postoperative total parenteral nutrition, critical. There is also a large fluid requirement over the first several days and chest tube outputs can be expected to be very high. Our routine is to ideally use colloid for these several days. Patients will also develop some degree of capillary leak syndrome and may become edematous. Concurrent with this fluid shift is a typical pattern of the operative lung being initially clear and subsequently “whiting out” over the first two to three days. This can occasionally involve the contralateral lung. Once the fluid demand begins to abate, diuresis is started.

We have used intraoperative PDT for numerous pleural cancers, both primary and metastatic. Anecdotally, we have seen some excellent results for metastatic breast and ovarian cancer, good results for thymic cancers, and generally poor results for primary or metastatic sarcomas. Our best quantified experience has been with NSCLC with pleural dissemination and mesothelioma.

NSCLC with pleural dissemination, stage IV (M1a), is a difficult clinical problem. We conducted a Phase II trial, when this disease was considered stage IIIB (AJCC 6th ed.). At that time there were no targeted chemotherapeutic agents and the median survival for this form of NSCLC was 6 to 9 months. According to our protocol, patients were treated with preoperative chemotherapy to “best effect” and then restaged. The oncologic eligibility criterion was disease confined to one hemithorax. N3 disease was an exclusionary, but N2 was not. In fact, 90% of the patients in the study had N2 disease. The operation the patients underwent was the appropriate anatomic resection for their primary lesion, which ranged from segmentectomy to pneumonectomy, and a parietal pleurectomy to achieve a macroscopic complete resection. Each patient then underwent PDT as described above with 2 mg/kg porfimer sodium administered 24 hours preoperatively to a measured dose of 30 J/cm² (using flat photodiodes which we subsequently correlated sided by side with the current isotropic detectors, which measured almost exactly double the amount of light, 60 J/cm²). Of the 22 patients enrolled, a macroscopic complete resection and PDT was accomplished in 17, 3 underwent incomplete resection and PDT, and 2 were unresectable. Local control at 6 months, the study objective was 73.3% in the 15 assessable patients. Median overall survival on an intention to treat analysis was 21.7 months. These survivals were calculated from the time of photosensitizer administration, that is, the day before surgery. The results were not calculated from the time of diagnosis or first treatment with chemotherapy (Table 124-1 and Fig. 124-9).²³

Overall, the results of this study were considered encouraging. Shortly after publication of this article ensued the era of targeted therapies. These therapies have demonstrated significant improvements in survival over previous treatment and hence, referral of patients abated. Subsequently, we have started to see patients who initially had excellent and durable responses to targeted therapies but have recurred or progressed. We have now operated on several of these patients and, anecdotally, are seeing benefit. Obviously, employing surgery for NSCLC remains investigational and should only be considered for highly selected patients under the auspices of a multidisciplinary tumor board in an Institutional Review Board approved protocol.

Malignant pleural mesothelioma remains the disease with which we have the greatest experience. We published a pilot study²² comparing the results of patients who underwent intraoperative PDT and either extrapleural pneumonectomy or radical pleurectomy. In our hands the radical pleurectomy patients had a significantly longer overall survival, despite little difference in disease-free survival. For that reason we switched exclusively to lung-sparing surgery. Subsequently, we reported our results for lung-sparing surgery and intraoperative PDT for 38 patients with malignant pleural mesothelioma.²⁴

The demographics and outcomes of this study are summarized in Table 124-2. A radical pleurectomy resulted in a macroscopic complete resection in 37 of the 38 patients. The one patient in whom a macroscopic complete resection could not be obtained, preoperative biopsy indicated epithelial subtype and a final pathology revealed a large desmoplastic component. Approximately, 20% of the patients had nonepithelial subtypes and 97% had stage III or IV cancer. There was a single postoperative mortality, a stroke. At a median follow-up of 34.4 months, the median survival was 31.7 months for all 38 patients, 41.2 months for the 31/38 epithelial patients, and 6.8 months for the nonepithelial patients. The median overall and progression-free survival for the 20/31 epithelial patients with N2 disease was 31.7 months and 15.1 months, respectively. The median survival for the epithelial patients with N1 or N0 disease was 57.1 months. The median progression-free survival was 9.6 months for all patients, 15.1 months for the epithelial patients, and 4.8 months for the nonepithelial patients. These results are illustrated in Fig. 124-10A– D. All of these survival statistics were calculated from the time of surgery, not the time of diagnosis or time of first treatment.

Bearing in mind the retrospective nature and limited size of this study, these results are very encouraging. Perhaps the most interesting observation was that the disease-free survival was not particularly long but the overall survival, especially for the epithelial subtype patients, was unusually long. This is notable within the context that 97% of these patients had stage III or IV disease and survival calculations were made from the time of surgery, not diagnosis or first treatment, which would have added months to their survival if that were the chosen beginning point. The reason for prolonged survival after recurrence is under investigation. One hypothesis is simply that having two lungs left the patients with more reserve to tolerate adjuvant treatments after recurrence. This is likely a contributing factor, though the survival results in this study were beyond other reported radical pleurectomy series. Given that fact, it suggests that the PDT may have played a role, as that is the distinguishing feature of this radical pleurectomy series. It is known that PDT can induce a tumor-specific immune reaction. Understanding that microscopic disease is left behind after any pleural surgery, and almost certainly more so with radical pleurectomy than extrapleural pneumonectomy, the possibility exists that the residual PDT-treated cells could induce an autologous tumor vaccine effect. This hypothesis is the focus of intense laboratory and clinical investigation at this time.

PDT is a unique and interesting cancer treatment. While it acts by a number of mechanisms (direct cell kill, selective destruction of neovasculature, nonspecific and specific immune effects), it can still be combined with essentially any other combination of cancer therapies. Herein, we have detailed the application of combining PDT with surgery for the treatment of pleural cancers. With proper dosimetry, PDT can be performed safely and in a straightforward and easy manner. The fact that PDT effect penetrates several millimeters below the surface is a feature of the treatment that emboldened our group to become increasingly aggressive with saving the lung, to the point where radical pleurectomy is currently our exclusive surgical approach for malignant pleural mesothelioma. With other cancers involving the pleura, such as NSCLC with pleural dissemination, we have shown that PDT can be safely combined with lung resection as well, ranging from segmentectomy to pneumonectomy. The results of using PDT for pleural cancers, especially mesothelioma, are encouraging. The possibility that a PDT-induced tumor-specific immune effect is contributing to the results we have observed is the focus of ongoing investigations and raises the tantalizing prospect of combining immunotherapies with the current approach to amplify this effect.

Surgical debulking of malignant mesothelioma can be accomplished with extrapleural pneumonectomy or radical pleurectomy. The frequent local recurrence within the same hemithorax, however, suggests that microscopic disease remains after these surgical procedures. Furthermore, the pattern of local recurrence preceding systemic recurrence suggests that any adjuvant therapy that kills the remaining microscopic cells is potentially curative. PDT is an innovative way to destroy cells several centimeters beyond the extrapleural plane of dissection. The authors clearly describe the details of this technique including many of the unforeseen pitfalls for the novice. They have worked out the details of light dosimetry and uniform light delivery. An open hemithorax after pleural resection lends itself particularly well to this technique. I share the authors hope that effective adjuvant therapy would preclude the need to remove the lung and diaphragm, thus providing a better quality of life and potential cure.